teradataml 17.20.0.6__py3-none-any.whl → 20.0.0.0__py3-none-any.whl

teradataml/automl/feature_engineering.py

@@ -0,0 +1,1648 @@
+# ################################################################## 
+# 
+# Copyright 2024 Teradata. All rights reserved.
+# TERADATA CONFIDENTIAL AND TRADE SECRET
+# 
+# Primary Owner: Sweta Shaw
+# Email Id: Sweta.Shaw@Teradata.com
+# 
+# Secondary Owner: Akhil Bisht
+# Email Id: AKHIL.BISHT@Teradata.com
+# 
+# Version: 1.1
+# Function Version: 1.0
+# ##################################################################
+
+# Python libraries
+import pandas as pd
+import time
+import json
+import re
+
+# Teradata libraries
+from teradataml.dataframe.dataframe import DataFrame
+from teradataml.dataframe.copy_to import copy_to_sql
+from teradataml import Antiselect
+from teradataml import BincodeFit, BincodeTransform
+from teradataml import ColumnSummary, CategoricalSummary, GetFutileColumns, FillRowId
+from teradataml import Fit, Transform
+from teradataml import NonLinearCombineFit, NonLinearCombineTransform
+from teradataml import NumApply
+from teradataml import OneHotEncodingFit, OneHotEncodingTransform
+from teradataml import OrdinalEncodingFit, OrdinalEncodingTransform
+from teradataml import SimpleImputeFit, SimpleImputeTransform
+from teradataml import StrApply
+from teradataml import TargetEncodingFit, TargetEncodingTransform
+from sqlalchemy import literal_column
+from teradatasqlalchemy import INTEGER
+from teradataml import display
+from teradataml.hyperparameter_tuner.utils import _ProgressBar
+from teradataml.utils.validators import _Validators
+
+
+class _FeatureEngineering:
+    
+    def __init__(self,
+                 data,
+                 target_column,
+                 model_list,
+                 verbose = 0,
+                 task_type = "Regression",
+                 custom_data = None):
+        """
+        DESCRIPTION:
+            Function initializes the data, target column and columns datatypes
+            for feature engineering.
+         
+        PARAMETERS:  
+            data:
+                Required Argument.
+                Specifies the input teradataml DataFrame for feature engineering.
+                Types: teradataml Dataframe
+            
+            target_column:
+                Required Arugment.
+                Specifies the name of the target column in "data"..
+                Types: str
+            
+            model_list:
+                Required Arugment.
+                Specifies the list of models to be used for model training.
+                Types: list
+            
+            verbose:
+                Optional Argument.
+                Specifies the detailed execution steps based on verbose level.
+                Default Value: 0
+                Permitted Values: 
+                    * 0: prints the progress bar and leaderboard
+                    * 1: prints the execution steps of AutoML.
+                    * 2: prints the intermediate data between the execution of each step of AutoML.
+                Types: int
+
+            task_type:
+                Required Arugment.
+                Specifies the task type for AutoML, whether to apply regresion OR classification
+                on the provived dataset.
+                Default Value: "Regression"
+                Permitted Values: "Regression", "Classification"
+                Types: str
+
+            custom_data:
+                Optional Arugment.
+                Specifies json object containing user customized input.
+                Types: json object
+        """
+        # Instance variables
+        self.data = data
+        self.target_column = target_column
+        self.model_list = model_list
+        self.verbose = verbose
+        self.task_type = task_type
+        self.custom_data = custom_data
+        self.excluded_cols=[]
+        self.data_types = {key: value for key, value in self.data._column_names_and_types} 
+        self.target_label = None
+        self.data_transform_dict = {}
+        self.one_hot_obj_count = 0
+        self.is_classification_type = lambda: self.task_type.upper() == 'CLASSIFICATION'
+    
+    # Method for doing feature engineering on data -> adding id, removing futile col, imputation, encoding(one hot)
+    def feature_engineering(self, 
+                            auto=True):
+        """
+        DESCRIPTION:
+            Function performs following operations :-
+                1. Removes futile columns/features from dataset.
+                2. Detects the columns with missing values.
+                3. Performs imputation on these columns with missing values.
+                4. Detects categorical columns and perform encoding on those columns.
+         
+        PARAMETERS:   
+            auto:
+                Optional Arugment.
+                Specifies whether to run AutoML in custom mode or auto mode.
+                When set to False, runs in custom mode. Otherwise, by default runs in auto mode.
+                Default Value: True
+                Types: boolean
+                
+        Returns:
+             tuple, First element represents teradataml DataFrame,
+             second element represents list of columns which are not participating in outlier tranformation.
+        """
+        # Assigning number of base jobs for progress bar.
+        base_jobs = 14 if auto else 18
+        
+        # Updating model list based on distinct value of target column for classification type
+        if self.is_classification_type():
+            if self.data.drop_duplicate(self.target_column).size > 2:
+                unsupported_models = ['svm', 'glm']
+                self.model_list = [model for model in self.model_list if model not in unsupported_models]
+
+        # Updating number of jobs for progress bar based on number of models.
+        jobs = base_jobs + len(self.model_list)
+        self.progress_bar = _ProgressBar(jobs=jobs, verbose=2, prefix='Automl Running:')
+
+        self._display_heading(phase=1, 
+                              progress_bar=self.progress_bar)
+        self._display_msg(msg='Feature Engineering started ...', 
+                          progress_bar=self.progress_bar)
+        
+        # Storing target column to data transform dictionary
+        self.data_transform_dict['data_target_column'] = self.target_column
+        # Storing target column encoding indicator to data transform dictionary
+        self.data_transform_dict['target_col_encode_ind'] = False
+        # Storing task type to data transform dictionary
+        self.data_transform_dict['classification_type']=self.is_classification_type()
+        # Storing params for performing one hot encoding
+        self.data_transform_dict['one_hot_encoding_fit_obj'] ={} 
+        self.data_transform_dict['one_hot_encoding_drop_list'] = []
+
+        if auto:
+            self._remove_duplicate_rows()
+            self.progress_bar.update()
+            
+            self._remove_futile_columns()
+            self.progress_bar.update()
+            
+            self._handle_date_columns()
+            self.progress_bar.update()
+            
+            self._handling_missing_value()
+            self.progress_bar.update()
+            
+            self._impute_missing_value()
+            self.progress_bar.update()
+            
+            self._encoding_categorical_columns()
+            self.progress_bar.update()
+            
+        else:
+            self._remove_duplicate_rows()
+            self.progress_bar.update()
+            
+            self._remove_futile_columns()
+            self.progress_bar.update()
+            
+            self._handle_date_columns()
+            self.progress_bar.update()
+            
+            self._custom_handling_missing_value()
+            self.progress_bar.update()
+            
+            self._bin_code_transformation()
+            self.progress_bar.update()
+            
+            self._string_manipulation()
+            self.progress_bar.update()
+            
+            self._custom_categorical_encoding()
+            self.progress_bar.update()
+            
+            self._mathematical_transformation()
+            self.progress_bar.update()
+            
+            self._non_linear_transformation()
+            self.progress_bar.update()
+            
+            self._anti_select_columns()
+            self.progress_bar.update()
+            
+        return self.data, self.excluded_cols, self.target_label, self.data_transform_dict
+    
+    def _extract_list(self,
+                      list1,
+                      list2):
+        """
+        DESCRIPTION:
+            Function to extract elements from list1 which are not present in list2.
+            
+        PARAMETERS:
+            list1:
+                Required Argument.
+                Specifies the first list for extracting elements from.
+                Types: list
+                
+            list2:
+                Required Argument.
+                Specifies the second list to get elements for avoiding in first list while extracting.
+                Types: list
+                
+        RETURN:
+            Returns extracted elements in form of list.
+            
+        """
+        new_lst = list(set(list1) - set(list2))
+        return new_lst
+    
+    def _remove_duplicate_rows(self):
+        """
+        DESCRIPTION:
+            Function to handles duplicate rows present in dataset.
+        
+        """
+        self._display_msg(msg="\nHandling duplicate records present in dataset ...",
+                          progress_bar=self.progress_bar,
+                          show_data=True)
+        start_time = time.time()
+        rows = self.data.shape[0]
+        self.data=self.data.drop_duplicate()
+        if rows != self.data.shape[0]:
+            self._display_msg(msg=f'Updated dataset sample after removing {rows-self.data.shape[0]} duplicate records:',
+                              data=self.data,
+                              progress_bar=self.progress_bar)
+            self._display_msg(inline_msg=f"Remaining Rows in the data: {self.data.shape[0]}\n"\
+                                  f"Remaining Columns in the data: {self.data.shape[1]}",
+                              progress_bar=self.progress_bar)
+        else:
+            self._display_msg(inline_msg="Analysis complete. No action taken.",
+                              progress_bar=self.progress_bar)   
+
+        end_time = time.time()
+        self._display_msg(msg="Total time to handle duplicate records: {:.2f} sec  ".format(end_time - start_time),
+                          progress_bar=self.progress_bar,
+                          show_data=True)
+        
+    def _get_distinct_count(self):
+        """
+        DESCRIPTION:
+            Function to get distinct count for all features and store it in dictionary for further use.
+        """
+        # Count of distinct value in each column
+        counts = self.data.select(self.data.columns).count(distinct=True)
+        
+        # Dict containing disctinct value in each column
+        self.counts_dict = next(counts.itertuples())._asdict()
+    
+    def _preprocess_data(self):
+        """
+        DESCRIPTION:
+            Function replaces the existing id column or adds the new id column and
+            removes columns with sinlge value/same values in the dataset.
+        """
+        # Get distinct value in each column
+        self._get_distinct_count()
+        
+        # Columns to removed if
+        # id column detected or count of distinct value = 1
+        columns_to_be_removed = [col for col in self.data.columns if col.lower() == 'id' or self.counts_dict[f'count_{col}'] == 1]
+
+        # Removing id column, if exists
+        if len(columns_to_be_removed) != 0:
+            self.data = self.data.drop(columns_to_be_removed, axis=1)
+            # Storing irrelevent column list in data transform dictionary
+            self.data_transform_dict['drop_irrelevent_columns'] = columns_to_be_removed
+            
+        # Adding id columns
+        obj = FillRowId(data=self.data, row_id_column='id')
+
+        self.data = obj.result
+            
+    def _remove_futile_columns(self):
+        """
+        DESCRIPTION:
+            Function removes the futile columns from dataset. 
+        """
+        self._display_msg(msg="\nHandling less significant features from data ...",
+                          progress_bar=self.progress_bar,
+                          show_data=True)
+        start_time = time.time()
+        
+        self._preprocess_data()
+        
+        # Handling string type target column in classification
+        # Performing Ordinal Encoding 
+        if self.data_types[self.target_column] in ['str']:
+            self._ordinal_encoding([self.target_column])
+    
+        # Detecting categorical columns
+        categorical_columns = [col for col, d_type in self.data._column_names_and_types if d_type == 'str']
+
+        # Detecting and removing futile columns, if categorical_column exists
+        if len(categorical_columns) != 0:
+            
+            obj = CategoricalSummary(data=self.data, 
+                                     target_columns=categorical_columns)
+            
+            gfc_out = GetFutileColumns(data=self.data, 
+                                       object=obj, 
+                                       category_summary_column="ColumnName",
+                                       threshold_value =0.7)
+            
+            # Extracting Futile columns
+            f_cols = [row[0] for row in gfc_out.result.itertuples()]
+            
+            if len(f_cols) == 0:
+                self._display_msg(inline_msg="All categorical columns seem to be significant.",
+                                  progress_bar=self.progress_bar)
+            else:
+
+                self.data = self.data.drop(f_cols, axis=1)
+                # Storing futile column list in data transform dictionary
+                self.data_transform_dict['futile_columns'] = f_cols
+                self._display_msg(msg='Removing Futile columns:',
+                                  col_lst=f_cols,
+                                  progress_bar=self.progress_bar)
+                self._display_msg(msg='Sample of Data after removing Futile columns:',
+                                  data=self.data,
+                                  progress_bar=self.progress_bar)
+        end_time= time.time()
+        self._display_msg(msg="Total time to handle less significant features: {:.2f} sec  ".format( end_time - start_time),
+                          progress_bar=self.progress_bar,
+                          show_data=True)
+
+    def _handle_date_component(self,
+                               date_component_columns,
+                               date_component):  
+        
+        """
+        DESCRIPTION:
+            Function to handle newly generated date components, i.e., day , month and year diff.
+            Based on their distinct values, binning is done with predefined prefix. 
+            Binned component is used further as categorical features.
+            
+        PARAMETERS:
+            date_component_columns:
+                Required Argument.
+                Specifies the list of newly generated differnt component of date features.
+                Types: list
+                
+            date_component:
+                Required Argument.
+                Specifies identifier for the differnt component of date features, i.e., D - Days , M - Months and Y - Year diffs.
+                Types: str 
+            
+        """
+        # Check for day
+        if date_component == "D":
+            prefix_value = "Day_"
+        # Check for month
+        elif date_component == "M":
+            prefix_value = "Month_"
+        # Check for year diff
+        elif date_component == "Y":
+            prefix_value = "Year_diff_"
+        
+        # Deciding bins based on distinct value of date component features.
+        for col in date_component_columns:
+            data_size = self.data.drop_duplicate(col).size
+            if data_size < 4:
+                num_bins = data_size
+            else:
+                num_bins = 4
+            # Performing bincode for converting date component to specific labels
+            fit_params = {
+                        "data": self.data,
+                        "target_columns": col,
+                        "method_type":"Equal-Width",
+                        "nbins": num_bins,
+                        "label_prefix" : prefix_value
+                        }
+            bin_code_fit = BincodeFit(**fit_params)
+            
+            fit_params_map = {"D": "day_component_fit_object",
+                              "M": "month_component_fit_object",
+                              "Y": "year_diff_component_fit_object"}
+            
+            # Storing fit object for each date component in data transform dictionary
+            self.data_transform_dict[fit_params_map[date_component]][col] = bin_code_fit.output
+                
+            accumulate_columns = self._extract_list(self.data.columns, [col])
+            transform_params = {
+                        "data": self.data,
+                        "object": bin_code_fit.output,
+                        "accumulate": accumulate_columns,
+                        "persist": True
+                        }
+            self.data = BincodeTransform(**transform_params).result
+          
+    def _fetch_date_component(self, 
+                              process, 
+                              regex_str, 
+                              columns, 
+                              date_component):
+        
+        """
+        DESCRIPTION:
+            Function to fetch newly generated date component features.
+            Passing ahead for performing binning.
+            
+        PARAMETERS:
+            process:
+                Required Argument.
+                Specifies date component of date feature which is going to be fetched and handled.
+                Types: str
+                
+            regex_str:
+                Required Argument.
+                Specifies regular expression for identifying newly generated date component features.
+                Types: str 
+            
+            columns:
+                Required Argument.
+                Specifies list of newly generated date component features.
+                Types: list
+                
+            date_component:
+                Required Argument.
+                Specifies identifier for the differnt component of date features, i.e., D - Days , M - Months and Y - Year diffs.
+                Types: str
+            
+        """
+        date_component_columns = [col for col in columns if re.search(regex_str+"$", col)]
+        if len(date_component_columns) != 0:
+            self._handle_date_component(date_component_columns,date_component)
+            self._display_msg(msg="Useful {} features:".format(process),
+                              col_lst=date_component_columns,
+                              progress_bar=self.progress_bar)
+            self._display_msg(msg="Updated dataset sample:",
+                              data=self.data,
+                              progress_bar=self.progress_bar)
+                
+        else:
+            self._display_msg("\nNo useful feature found for {} component:".format(process),
+                              progress_bar=self.progress_bar)
+            
+        return date_component_columns
+  
+    def _handle_date_columns_helper(self):
+        
+        """
+        DESCRIPTION:
+            Function for dropping irrelevent date features.
+            Extracting day, month and year component from revelent date features.
+            Passing extracted component for performing binning.
+        """
+        
+        # Dropping missing value for all date columns
+        self._display_msg(msg="\nDropping missing values for:",
+                          col_lst=self.date_column_list,
+                          progress_bar=self.progress_bar)
+
+        self.data = self.data.dropna(subset=self.date_column_list)
+     
+        # Date columns list eligible for dropping from dataset
+        drop_date_cols = []     
+        
+        # Checking for single valued date columns
+        for col in self.date_column_list:
+            if self.data.drop_duplicate(col).size == self.data.shape[0]:
+                drop_date_cols.append(col)
+                
+        if len(drop_date_cols) != 0:
+            self.data = self.data.drop(drop_date_cols, axis=1)
+            # Storing unique date column list in data transform dictionary
+            self.data_transform_dict['drop_unique_date_columns'] = drop_date_cols
+            self._display_msg(msg='Dropping date features with all unique value:',
+                              col_lst = drop_date_cols,
+                              progress_bar=self.progress_bar)
+             
+        # Updated date columns list
+        self.date_column_list = [item for item in self.date_column_list if item not in drop_date_cols]
+        
+        # List for storing newly generated date component features
+        new_columns=[]
+        
+        # Extracting day, month and year difference from date columns
+        if len(self.date_column_list) != 0:
+            
+            component_param={}
+            for col in self.date_column_list: 
+                
+                day_column=str(col)+"_day_comp"
+                month_column=str(col)+"_month_comp"
+                year_diff_column=str(col)+"_year_diff_comp"
+                new_columns.extend([day_column,month_column,year_diff_column])
+                day_query=("EXTRACT(DAY FROM {0})".format(col))
+                month_query=("EXTRACT(MONTH FROM {0})".format(col))
+                year_query=("EXTRACT(YEAR FROM CURRENT_DATE) - EXTRACT(YEAR FROM {0})".format(col))
+                component_param[day_column]=literal_column(day_query,INTEGER())
+                component_param[month_column]=literal_column(month_query,INTEGER())
+                component_param[year_diff_column]=literal_column(year_query,INTEGER())
+                
+            self.data=self.data.assign(**component_param)
+            # Storing newly generated date component list along with parameters in data transform dictionary
+            self.data_transform_dict['extract_date_comp_col'] = self.date_column_list
+            self.data_transform_dict['extract_date_comp_param'] = component_param
+         
+            # Dropping date columns as we have already extracted day, month and year in new columns        
+            self.data = self.data.drop(self.date_column_list, axis=1)
+            self._display_msg(msg='List of newly generated features from existing date features:',
+                              col_lst=new_columns,
+                              progress_bar=self.progress_bar)
+            self._display_msg(msg='List of newly generated features from existing date features:',
+                              data=self.data,
+                              progress_bar=self.progress_bar)
+            
+            drop_cols=[]
+            
+            for col in new_columns:
+                distinct_rows = self.data.drop_duplicate(col).size
+                if  distinct_rows == self.data.shape[0]:
+                    drop_cols.append(col)
+                    self._display_msg(msg='Dropping features with all unique values:',
+                                      col_lst=col,
+                                      progress_bar=self.progress_bar)
+                    
+                elif distinct_rows == 1:
+                    drop_cols.append(col)
+                    self._display_msg(msg='Dropping features with single value:',
+                                      col_lst=col,
+                                      progress_bar=self.progress_bar)
+
+            # Dropping columns from drop_cols list
+            if len(drop_cols) != 0:
+                self.data = self.data.drop(drop_cols, axis=1)
+                # Storing extract date component list for drop in data transform dictionary
+                self.data_transform_dict['drop_extract_date_columns'] = drop_cols        
+                
+            # Extracting all newly generated columns    
+            new_columns = [item for item in new_columns if item not in drop_cols]
+            
+            # Storing each date component transformation fit object in data transform dictionary
+            self.data_transform_dict = {**self.data_transform_dict, 
+                                        'day_component_fit_object': {}, 
+                                        'month_component_fit_object': {}, 
+                                        'year_diff_component_fit_object': {}}
+            # Grouping date components based on types i.e., day, month, and year_diff for performing binning
+            if len(new_columns) != 0:
+                self.day_columns = self._fetch_date_component("day", "_day_comp", new_columns, "D")
+                self.month_columns = self._fetch_date_component("month", "_month_comp", new_columns, "M")
+                self.year_diff_columns = self._fetch_date_component("year_diff", "_year_diff_comp", new_columns, "Y")
+                self._display_msg(inline_msg="No useful date component found",
+                                  progress_bar=self.progress_bar)
+            
+            self._display_msg(msg='Updated dataset sample after handling date features:',
+                              data=self.data,
+                              progress_bar=self.progress_bar)
+        else:
+            self._display_msg(inline_msg="No useful date feature found",
+                              progress_bar=self.progress_bar)
+             
+    def _handle_date_columns(self):
+        
+        """
+        DESCRIPTION:
+            Function to handle date columns in dataset if any. 
+            Perform relevent transformation by extracting different components, i.e., Day , Month and Year.
+        """
+        self._display_msg(msg="\nHandling Date Features ...",
+                          progress_bar=self.progress_bar,
+                          show_data=True)
+        start_time = time.time()
+        
+        self.date_column_list = [col for col, d_type in self.data._column_names_and_types \
+                               if d_type in ["datetime.date","datetime.datetime"]]
+                 
+        if len(self.date_column_list) == 0:
+            self._display_msg(inline_msg="Dataset does not contain any feature related to dates.",
+                              progress_bar=self.progress_bar)
+        else:
+            # Storing date column list in data transform dictionary
+            self.data_transform_dict['date_columns'] = self.date_column_list
+            self._handle_date_columns_helper()
+        
+        end_time = time.time()
+        self._display_msg(msg="Total time to handle date features: {:.2f} sec\n".format(end_time-start_time),
+                          progress_bar=self.progress_bar,
+                          show_data=True)         
+
+    def _missing_count_per_column(self):
+        """
+        DESCRIPTION:
+            Function finds and returns a dictnoary containing list of columns
+            with missing values.
+         
+        Returns:
+            dict, keys represent column names and
+            values represent the missing value count for corresponding column.
+        """
+        
+        # Removing rows with missing target column value
+        self.data = self.data.dropna(subset=[self.target_column])
+        
+        obj = ColumnSummary(data=self.data, 
+                            target_columns=self.data.columns, 
+                            volatile=True)
+        
+        cols_miss_val={}
+        # Iterating over each row in the column summary result
+        for row in obj.result.itertuples():
+            # Checking if the third element of the row (missing values count) is greater than 0
+            if row[3] > 0:
+                # If so, add an entry to the 'cols_miss_val' dictionary
+                # Key: column name (first element of the row)
+                # Value: count of missing values in the column (third element of the row)
+                cols_miss_val[row[0]] = row[3]
+        
+        return cols_miss_val
+ 
+    def _handling_missing_value(self):
+        """
+        DESCRIPTION:
+            Function detects the missing values in the each feature of dataset,
+            then performs these operation based on condition :-
+                1. deleting rows from columns/feature
+                2. dropping columns from dataset  
+        """
+        self._display_msg(msg="\nChecking Missing values in dataset ...",
+                          progress_bar=self.progress_bar,
+                          show_data=True)
+        start_time = time.time()
+        
+        # Flag for missing values
+        msg_val_found=0
+        
+        #num of rows
+        d_size = self.data.shape[0]
+
+        delete_rows = []
+        drop_cols = []
+        self.imputation_cols = {}
+
+        cols_miss_val = self._missing_count_per_column()
+
+        if len(cols_miss_val) != 0:
+            self._display_msg(msg="Columns with their missing values:",
+                              col_lst=cols_miss_val,
+                              progress_bar=self.progress_bar)
+        
+        # Get distinct value in each column
+        self._get_distinct_count()
+        
+        # Iterating over columns with missing values
+        for col,val in  cols_miss_val.items():
+            
+            # Drop col, if count of missing value > 60%
+            if val > .6*d_size:
+                drop_cols.append(col)
+                continue
+            
+            if self.data_types[col] in ['float', 'int']:
+                corr_df = self.data[col].corr(self.data[self.target_column])
+                corr_val = self.data.assign(True, corr_=corr_df)
+                related = next(corr_val.itertuples())[0]
+                
+                # Delete row, if count of missing value < 2% and 
+                # Relation b/w target column and numeric column <= .25
+                if val < .02*d_size and related <= .25:
+                    delete_rows.append(col)
+                    continue
+                    
+            elif self.data_types[col] in ['str']:
+                # Delete row, if count of missing value < 4%
+                if val < .04*d_size:
+                    delete_rows.append(col)
+                    continue
+                # Drop col, if unique count of column > 75%
+                elif self.counts_dict[f'count_{col}'] > .75*(d_size-val):
+                    drop_cols.append(col)
+                    continue
+                    
+            # Remaining column for imputation
+            self.imputation_cols[col] = val
+            # Storing columns with missing value for imputation in data transform dictionary
+            self.data_transform_dict['imputation_columns'] = self.imputation_cols
+
+        if len(delete_rows) != 0:
+            self.data = self.data.dropna(subset=delete_rows)
+            msg_val_found=1
+            self._display_msg(msg='Deleting rows of these columns for handling missing values:',
+                              col_lst=delete_rows,
+                              progress_bar=self.progress_bar)
+            
+        if len(drop_cols) != 0:
+            self.data = self.data.drop(drop_cols, axis=1)
+            msg_val_found=1
+            # Storing columns with missing value for drop in data transform dictionary
+            self.data_transform_dict['drop_missing_columns'] = drop_cols
+            self._display_msg(msg='Dropping these columns for handling missing values:',
+                              col_lst=drop_cols,
+                              progress_bar=self.progress_bar)
+        
+        if len(self.imputation_cols) == 0 and msg_val_found ==0:
+            self._display_msg(inline_msg="No Missing Values Detected.",
+                              progress_bar=self.progress_bar)
+            
+        end_time = time.time()
+        self._display_msg(msg="Total time to find missing values in data: {:.2f} sec  ".format( end_time - start_time),
+                          progress_bar=self.progress_bar,
+                          show_data=True)
+
+    def _impute_helper(self):
+        """
+        DESCRIPTION:
+            Function decides the imputation methods [mean/ median/ mode] for columns with missing values
+            on the basis of skewness of column in the dataset.
+        
+        RETURNS:
+            A tuple containing,
+            col_stat (name of columns with missing value)
+            stat (imputation method for respective columns)
+        """
+        col_stat = []
+        stat = []
+        
+        # Converting o/p of skew() into dictonary with key as column name and value as skewness value
+        df = self.data.skew()
+        skew_data = next(df.itertuples())._asdict()
+        
+        # Iterating over columns with missing value
+        for key, val in self.imputation_cols.items():
+            
+            col_stat.append(key)
+            if self.data_types[key] in ['float', 'int']:
+                val = skew_data[f'skew_{key}']
+                # Median imputation method, if abs(skewness value) > 1
+                if abs(val) > 1:
+                    stat.append('median')
+                # Mean imputation method, if abs(skewness value) <= 1
+                else:
+                    stat.append('mean')     
+            # Mode imputation method, if categorical column
+            else:
+                stat.append('mode')
+        
+        self._display_msg(msg="Columns with their imputation method:",
+                          col_lst=dict(zip(col_stat, stat)),
+                          progress_bar=self.progress_bar)
+        
+        return col_stat, stat
+
+    def _impute_missing_value(self):
+        """
+        DESCRIPTION:
+            Function performs the imputation on columns/features with missing values in the dataset.
+        """
+        
+        start_time = time.time()
+        self._display_msg(msg="\nImputing Missing Values ...",
+                          progress_bar=self.progress_bar,
+                          show_data=True)
+        
+        if len(self.imputation_cols) != 0:
+            
+            # List of columns and imputation Method
+            col_stat, stat = self._impute_helper()
+            
+            fit_obj = SimpleImputeFit(data=self.data, 
+                                      stats_columns=col_stat, 
+                                      stats=stat, 
+                                      volatile=True)
+            
+            # Storing fit object for imputation in data transform dictionary
+            self.data_transform_dict['imputation_fit_object'] = fit_obj.output
+            sm = SimpleImputeTransform(data=self.data, 
+                                        object=fit_obj, 
+                                        volatile=True)
+            
+            self.data = sm.result
+            self._display_msg(msg="Sample of Data after Imputation:",
+                              data=self.data,
+                              progress_bar=self.progress_bar)
+        else:
+            self._display_msg(inline_msg="No imputation is Required.",
+                              progress_bar=self.progress_bar)
+              
+        end_time = time.time()
+        self._display_msg(msg="Time taken to perform imputation: {:.2f} sec  ".format(end_time - start_time),
+                          progress_bar=self.progress_bar,
+                          show_data=True)
+
+    
+    def _custom_handling_missing_value(self):
+        """
+        DESCRIPTION:
+            Function to perform customized missing value handling for features based on user input.
+            
+        """
+        # Fetching user input for performing missing value handling
+        missing_handling_input = self.custom_data.get("MissingValueHandlingIndicator", False)
+        
+        if missing_handling_input:
+            # Fetching parameters required for performing
+            missing_handling_param = self.custom_data.get("MissingValueHandlingParam", None)
+            if missing_handling_param:
+                # Fetching user input for different methods missing value handling
+                drop_col_ind = missing_handling_param.get("DroppingColumnIndicator", False)
+                drop_row_ind = missing_handling_param.get("DroppingRowIndicator", False)
+                impute_ind = missing_handling_param.get("ImputeMissingIndicator", False)
+                # Checking for user input if all methods indicator are false or not
+                if not any([drop_col_ind, drop_row_ind, impute_ind]):  
+                    self._display_msg(inline_msg="No method information provided for performing customized missing value handling. \
+                        AutoML will proceed with default missing value handling method.",
+                                      progress_bar=self.progress_bar)
+                    
+                else:
+                    # Checking user input for dropping missing value columns
+                    if drop_col_ind:
+                        drop_col_list = missing_handling_param.get("DroppingColumnList", [])
+                        # Storing customcolumns with missing value for drop in data transform dictionary
+                        self.data_transform_dict["custom_drop_missing_columns"] = drop_col_list
+                        if len(drop_col_list):
+                            # Checking for column present in dataset or not
+                            _Validators._validate_dataframe_has_argument_columns(drop_col_list, "DroppingColumnList", self.data, "df")
+
+                            self._display_msg(msg="\nDropping these columns for handling customized missing value:",
+                                              col_lst=drop_col_list,
+                                              progress_bar=self.progress_bar)
+                            self.data = self.data.drop(drop_col_list, axis=1)
+                        else:
+                            self._display_msg(inline_msg="No information provided for dropping missing value containing columns.",
+                                              progress_bar=self.progress_bar)
+
+                    # Checking user input for dropping missing value rows    
+                    if drop_row_ind:
+                        drop_row_list = missing_handling_param.get("DroppingRowList", [])
+                        if len(drop_row_list):
+                            # Checking for column present in dataset or not
+                            _Validators._validate_dataframe_has_argument_columns(drop_row_list, "DroppingRowList", self.data, "df")
+
+                            self._display_msg(msg="Dropping missing rows in these columns for handling customized missing value:",
+                                              col_lst=drop_row_list,
+                                              progress_bar=self.progress_bar)
+                            self.data = self.data.dropna(subset = drop_row_list)
+                        else:
+                            self._display_msg(inline_msg="No information provided for dropping missing value containing rows.",
+                                              progress_bar=self.progress_bar)
+                    # Checking user input for missing value imputation 
+                    if impute_ind:
+                        stat_list = missing_handling_param.get("StatImputeList", None)
+                        stat_method = missing_handling_param.get("StatImputeMethod", None)
+                        literal_list = missing_handling_param.get("LiteralImputeList", None)
+                        literal_value = missing_handling_param.get("LiteralImputeValue", None)
+
+                        # Checking for column present in dataset or not
+                        _Validators._validate_dataframe_has_argument_columns(stat_list, "StatImputeList", self.data, "df")
+
+                        _Validators._validate_dataframe_has_argument_columns(literal_list, "LiteralImputeList", self.data, "df")
+                        
+                        # Creating fit params    
+                        fit_param = {
+                            "data" : self.data,
+                            "stats_columns" : stat_list,
+                            "stats" : stat_method,
+                            "literals_columns" : literal_list,
+                            "literals" : literal_value
+                        }
+                        # Fitting on dataset
+                        fit_obj = SimpleImputeFit(**fit_param)
+                        # Storing custom fit object for imputation in data transform dictionary
+                        self.data_transform_dict["custom_imputation_ind"] = True
+                        self.data_transform_dict["custom_imputation_fit_object"] = fit_obj.output
+                        # Creating transform params
+                        transform_param = {
+                            "data" : self.data,
+                            "object" : fit_obj.output,
+                            "persist" : True
+                        }
+                        # Updating dataset with transform result
+                        self.data = SimpleImputeTransform(**transform_param).result
+                        self._display_msg(msg="Updated dataset sample after performing customized missing value imputation:",
+                                          data=self.data,
+                                          progress_bar=self.progress_bar)
+            else:
+                self._display_msg(inline_msg="No information provided for performing customized missing value handling. \
+                        AutoML will proceed with default missing value handling method.",
+                                  progress_bar=self.progress_bar)
+        else:
+            self._display_msg(inline_msg="Proceeding with default option for missing value imputation.",
+                              progress_bar=self.progress_bar)
+            
+        # Proceeding with default method for handling remaining missing values
+        self._display_msg(inline_msg="Proceeding with default option for handling remaining missing values.",
+                          progress_bar=self.progress_bar)
+        self._handling_missing_value()
+        self._impute_missing_value()   
+
+    def _bin_code_transformation(self):
+        """
+        DESCRIPTION:
+            Function to perform customized binning on features based on user input.
+            
+        """
+        # Fetching user input for performing bin code transformation.
+        bin_code_input = self.custom_data.get("BincodeIndicator", False)
+         
+        if bin_code_input:
+            # Storing custom bin code transformation indicator in data transform dictionary
+            self.data_transform_dict['custom_bincode_ind'] = True
+            # Fetching list required for performing transfomation.
+            extracted_col = self.custom_data.get("BincodeParam", None)
+            if not extracted_col:
+                self._display_msg(inline_msg="BincodeParam is empty. Skipping customized bincode transformation.",
+                                  progress_bar=self.progress_bar)
+            else:
+                # Creating list for storing column and binning informartion for performing transformation 
+                equal_width_bin_list  = []
+                equal_width_bin_columns  = []
+                var_width_bin_list = []
+                var_width_bin_columns  = []
+
+                # Checking for column present in dataset or not
+                _Validators._validate_dataframe_has_argument_columns(list(extracted_col.keys()), "BincodeParam", self.data, "df")
+                
+                for col,transform_val in extracted_col.items():
+                    # Fetching type of binning to be performed 
+                    bin_trans_type = transform_val["Type"]
+                    # Fetching number of bins to be created
+                    num_bin = transform_val["NumOfBins"]
+                    # Checking for bin types and adding details into lists for binning
+                    if bin_trans_type == "Equal-Width":
+                        bins = num_bin
+                        equal_width_bin_list.append(bins)
+                        equal_width_bin_columns.append(col)
+                    elif bin_trans_type == "Variable-Width":
+                        var_width_bin_columns.append(col)
+                        bins = num_bin
+                        for i in range(1, bins+1):
+                            # Forming binning name as per expected input
+                            temp="Bin_"+str(i)
+                            # Fetching required details for variable type binning
+                            minval = transform_val[temp]["min_value"]
+                            maxval = transform_val[temp]["max_value"]
+                            label = transform_val[temp]["label"]
+                            # Appending information of each bin             
+                            var_width_bin_list.append({ "ColumnName":col, "MinValue":minval, "MaxValue":maxval, "Label":label})
+                # Checking column list for performing binning with Equal-Width.
+                if len(equal_width_bin_columns) != 0:
+                    # Adding fit parameter for performing binning with Equal-Width.
+                    fit_params={
+                        "data" : self.data,
+                        "target_columns": equal_width_bin_columns,
+                        "method_type" : "Equal-Width",
+                        "nbins" : bins  
+                    }
+                    eql_bin_code_fit = BincodeFit(**fit_params)
+                    # Storing fit object and column list for Equal-Width binning in data transform dictionary
+                    self.data_transform_dict['custom_eql_bincode_col'] = equal_width_bin_columns
+                    self.data_transform_dict['custom_eql_bincode_fit_object'] = eql_bin_code_fit.output
+                    # Extracting accumulate columns
+                    accumulate_columns = self._extract_list(self.data.columns, equal_width_bin_columns)
+                    # Adding transform parameters for performing binning with Equal-Width.
+                    eql_transform_params={
+                        "data" : self.data,
+                        "object" : eql_bin_code_fit.output,
+                        "accumulate" : accumulate_columns,
+                        "persist" : True,   
+                    }
+                    self.data = BincodeTransform(**eql_transform_params).result
+                    self._display_msg(msg="\nUpdated dataset sample after performing Equal-Width binning :-",
+                                      data=self.data,
+                                      progress_bar=self.progress_bar)
+                else:
+                    self._display_msg(inline_msg="No information provided for Equal-Width Transformation.",
+                                      progress_bar=self.progress_bar)
+                    
+                if len(var_width_bin_columns) != 0:
+                    # Creating pandas dataframe and then teradata dataframe for storing binning information
+                    var_bin_table = pd.DataFrame(var_width_bin_list, columns=["ColumnName", "MinValue", "MaxValue", "Label"])
+                    self._display_msg(msg="Variable-Width binning information:-",
+                                      data=var_bin_table,
+                                      progress_bar=self.progress_bar)
+                    copy_to_sql(df=var_bin_table, table_name="automl_bincode_var_fit", temporary=True)
+                    var_fit_input = DataFrame.from_table("automl_bincode_var_fit")
+                    fit_params = {
+                        "data" : self.data,
+                        "fit_data": var_fit_input,
+                        "fit_data_order_column" : ["MinValue", "MaxValue"],
+                        "target_columns": var_width_bin_columns,
+                        "minvalue_column" : "MinValue",
+                        "maxvalue_column" : "MaxValue",
+                        "label_column" : "Label",
+                        "method_type" : "Variable-Width",
+                        "label_prefix" : "label_prefix"
+                    }
+                    var_bin_code_fit = BincodeFit(**fit_params)
+                    # Storing fit object and column list for Variable-Width binning in data transform dictionary
+                    self.data_transform_dict['custom_var_bincode_col'] = var_width_bin_columns
+                    self.data_transform_dict['custom_var_bincode_fit_object'] = var_bin_code_fit.output
+                    accumulate_columns = self._extract_list(self.data.columns, var_width_bin_columns)
+                    var_transform_params = {
+                        "data" : self.data,
+                        "object" : var_bin_code_fit.output,
+                        "object_order_column" : "TD_MinValue_BINFIT",
+                        "accumulate" : accumulate_columns,
+                        "persist" : True   
+                    }
+                    self.data = BincodeTransform(**var_transform_params).result
+                    self._display_msg(msg="Updated dataset sample after performing Variable-Width binning:",
+                                      data=self.data,
+                                      progress_bar=self.progress_bar)
+                else:
+                    self._display_msg(inline_msg="No information provided for Variable-Width Transformation.",
+                                      progress_bar=self.progress_bar)            
+        else:
+            self._display_msg(inline_msg="No information provided for Variable-Width Transformation.",
+                              progress_bar=self.progress_bar)
+
+    def _string_manipulation(self):
+        """
+        DESCRIPTION:
+            Function to perform customized string manipulations on categorical features based on user input.
+                    
+        """
+        # Fetching user input for performing string manipulation.
+        str_mnpl_input = self.custom_data.get("StringManipulationIndicator", False)
+        # Checking user input for string manipulation on categrical features.
+        if str_mnpl_input:
+            # Storing custom string manipulation indicator in data transform dictionary
+            self.data_transform_dict['custom_string_manipulation_ind'] = True
+            # Fetching list required for performing operation.
+            extracted_col = self.custom_data.get("StringManipulationParam", None)
+            if not extracted_col:
+                self._display_msg(inline_msg="No information provided for performing string manipulation.",
+                                  progress_bar=self.progress_bar)
+            else:
+                # Checking for column present in dataset or not
+                _Validators._validate_dataframe_has_argument_columns(list(extracted_col.keys()), "StringManipulationParam", self.data, "df")
+
+                for target_col,transform_val in extracted_col.items():                
+                    self.data = self._str_method_mapping(target_col, transform_val)
+                # Storing custom string manipulation parameters in data transform dictionary
+                self.data_transform_dict['custom_string_manipulation_param'] = extracted_col
+                
+                self._display_msg(msg="Updated dataset sample after performing string manipulation:",
+                                  data=self.data,
+                                  progress_bar=self.progress_bar)
+        else:
+            self._display_msg(inline_msg="Skipping customized string manipulation.")
+                     
+    def _str_method_mapping(self,
+                            target_col,
+                            transform_val):
+        """
+        DESCRIPTION:
+            Function to map customized parameters according to passed method and 
+            performs string manipulation on categorical features.
+            
+        PARAMETERS:
+            target_col:
+                Required Argument.
+                Specifies feature for applying string manipulation.
+                Types: str
+                
+            transform_val:
+                Required Argument.
+                Specifies different parameter require for applying string manipulation.
+                Types: dict
+        
+        RETURNS:
+                Dataframe containing transformed data after applying string manipulation.
+        
+        """
+        # Creating list of features for accumulating while performing string manipulation on certain features
+        accumulate_columns = self._extract_list(self.data.columns, [target_col])
+        
+        # Fetching required parameters from json object
+        string_operation = transform_val["StringOperation"]
+        
+        # Storing general parameters for performing string transformation
+        fit_params = {
+            "data" : self.data,
+            "target_columns" : target_col,
+            "string_operation" : string_operation,
+            "accumulate" : accumulate_columns,
+            "inplace" : True,
+            "persist" : True
+        }
+        # Adding additional parameters based on string operation type           
+        if string_operation in ["StringCon", "StringTrim"]:
+            string_argument = transform_val["String"] 
+            fit_params = {**fit_params, 
+                          "string" : string_argument}
+        elif string_operation == "StringPad":
+            string_argument = transform_val["String"]
+            string_length = transform_val["StringLength"]
+            fit_params = {**fit_params, 
+                          "string" : string_argument, 
+                          "string_length" : string_length}       
+        elif string_operation == "Substring":
+            string_index = transform_val["StartIndex"]
+            string_length = transform_val["StringLength"]
+            fit_params = {**fit_params, 
+                          "start_index" : string_index,
+                          "string_length" : string_length}
+        
+        # returning dataset after performing string manipulation                 
+        return StrApply(**fit_params).result 
+              
+    def _one_hot_encoding(self, 
+                         one_hot_columns, 
+                         unique_counts):
+        """
+        DESCRIPTION:
+            Function performs the one hot encoding to categorcial columns/features in the dataset.
+         
+        PARAMETERS:
+            one_hot_columns: 
+                  Required Argument.
+                  Specifies the categorical columns for which one hot encoding will be performed.
+                  Types: str or list of strings (str)
+            
+            unique_counts:
+                  Required Argument.
+                  Specifies the unique counts in the categorical columns.
+                  Types: int or list of integer (int)
+                  
+        """
+        # TD function will add extra column_other in onehotEncoding, so 
+        # initailizing this list to remove those extra columns 
+        drop_lst = [ele + "_other" for ele in one_hot_columns]
+        # Adding fit parameters for performing encoding
+        fit_params = {
+            "data" : self.data,
+            "approach" : "auto",
+            "is_input_dense" : True,
+            "target_column" : one_hot_columns,
+            "category_counts" : unique_counts,
+            "other_column" : "other"
+        }
+        # Performing one hot encoding fit on target columns
+        fit_obj = OneHotEncodingFit(**fit_params)
+        # Storing indicator, fit object and column drop list for one hot encoding in data transform dictionary
+        self.data_transform_dict['one_hot_encoding_ind'] = True
+        self.data_transform_dict['one_hot_encoding_fit_obj'].update({self.one_hot_obj_count : fit_obj.result})
+        self.data_transform_dict['one_hot_encoding_drop_list'].extend(drop_lst)
+        self.one_hot_obj_count = self.one_hot_obj_count + 1
+        # Adding transform parameters for performing encoding
+        transform_params = {
+            "data" : self.data, 
+            "object" : fit_obj.result, 
+            "is_input_dense" : True,
+            "persist" : True
+        }
+        # Performing one hot encoding transformation
+        transform_obj = OneHotEncodingTransform(**transform_params)
+        self.data = transform_obj.result.drop(drop_lst, axis=1)
+
+    def _ordinal_encoding(self,
+                          ordinal_columns):
+        """
+        DESCRIPTION:
+            Function performs the ordinal encoding to categorcial columns or features in the dataset.
+         
+        PARAMETERS:
+            ordinal_columns: 
+                Required Argument.
+                Specifies the categorical columns for which ordinal encoding will be performed.
+                Types: str or list of strings (str)
+        """
+        # Adding fit parameters for performing encoding
+        fit_params = {
+            "data" : self.data,
+            "target_column" : ordinal_columns,
+            "volatile" : True
+        }
+        # Performing ordinal encoding fit on target columns
+        ord_fit_obj = OrdinalEncodingFit(**fit_params)
+        # Storing fit object and column list for ordinal encoding in data transform dictionary
+        if ordinal_columns[0] != self.target_column:
+            self.data_transform_dict["custom_ord_encoding_fit_obj"] = ord_fit_obj.result
+            self.data_transform_dict['custom_ord_encoding_col'] = ordinal_columns
+        else:
+            self.data_transform_dict['target_col_encode_ind'] = True
+            self.data_transform_dict['target_col_ord_encoding_fit_obj'] = ord_fit_obj.result
+        # Extracting accumulate columns
+        accumulate_columns = self._extract_list(self.data.columns, ordinal_columns)
+        # Adding transform parameters for performing encoding
+        transform_params = {
+            "data" : self.data,
+            "object" : ord_fit_obj.result,
+            "accumulate" : accumulate_columns,
+            "persist" : True
+        }
+        # Performing ordinal encoding transformation
+        self.data = OrdinalEncodingTransform(**transform_params).result
+        
+        if len(ordinal_columns) == 1 and ordinal_columns[0] == self.target_column:
+            self.target_label = ord_fit_obj
+        
+
+    def _target_encoding(self,
+                          target_encoding_list):
+        """
+        DESCRIPTION:
+            Function performs the target encoding to categorcial columns/features in the dataset.
+         
+        PARAMETERS:
+            target_encoding_list: 
+                  Required Argument.
+                  Specifies the categorical columns for which target encoding will be performed.
+                  Types: str or list of strings (str)
+        """
+        # Fetching all columns on which target encoding will be performed.
+        target_columns= list(target_encoding_list.keys())
+        # Checking for column present in dataset or not
+        _Validators._validate_dataframe_has_argument_columns(target_columns, "TargetEncodingList", self.data, "df")
+        # Finding distinct values and counts for columns.
+        cat_sum = CategoricalSummary(data = self.data,
+                                    target_columns = target_columns)
+        category_data=cat_sum.result.groupby("ColumnName").count()
+        category_data = category_data.assign(drop_columns = True,
+                                            ColumnName = category_data.ColumnName,
+                                            CategoryCount = category_data.count_DistinctValue)
+        # Storing indicator and fit object for target encoding in data transform dictionary
+        self.data_transform_dict["custom_target_encoding_ind"] = True
+        self.data_transform_dict["custom_target_encoding_fit_obj"] = {}
+        # Fetching required argument for performing target encoding
+        for col,transform_val in target_encoding_list.items():
+            encoder_method = transform_val["encoder_method"]
+            response_column = transform_val["response_column"]
+            # Adding fit parameters for performing encoding
+            fit_params = {
+                "data" : self.data,
+                "category_data" : category_data,
+                "encoder_method" : encoder_method,
+                "target_columns" : col,
+                "response_column" : response_column
+                }
+            if encoder_method == "CBM_DIRICHLET":
+                num_distinct_responses=transform_val["num_distinct_responses"]
+                fit_params = {**fit_params, 
+                            "num_distinct_responses" : num_distinct_responses}
+            # Performing target encoding fit on target columns
+            tar_fit_obj = TargetEncodingFit(**fit_params)
+            # Storing each column fit object for target encoding in data transform dictionary
+            self.data_transform_dict["custom_target_encoding_fit_obj"].update({col : tar_fit_obj})
+            # Extracting accumulate columns
+            accumulate_columns = self._extract_list(self.data.columns, [col])
+            # Adding transform parameters for performing encoding
+            transform_params = {
+                "data" : self.data,
+                "object" : tar_fit_obj,
+                "accumulate" : accumulate_columns,
+                "persist" : True                
+            }
+            # Performing ordinal encoding transformation
+            self.data = TargetEncodingTransform(**transform_params).result
+    
+    def _encoding_categorical_columns(self):
+        """
+        DESCRIPTION:
+            Function detects the categorical columns and performs encoding on categorical columns in the dataset.
+        """
+        self._display_msg(msg="\nPerforming encoding for categorical columns ...",
+                          progress_bar=self.progress_bar,
+                          show_data=True)
+        start_time = time.time()
+        
+        ohe_col = []
+        unique_count  = []
+
+        # List of columns before one hot
+        col_bf_ohe = self.data.columns
+        
+        # Get distinct value in each column
+        self._get_distinct_count()
+        
+        # Detecting categorical columns with thier unique counts
+        for col, d_type in self.data._column_names_and_types:
+            if d_type in ['str']:
+                ohe_col.append(col)
+                unique_count.append(self.counts_dict[f'count_{col}'])
+
+        if len(ohe_col) != 0:
+            self._one_hot_encoding(ohe_col, unique_count)
+
+            self._display_msg(msg="ONE HOT Encoding these Columns:",
+                              col_lst=ohe_col,
+                              progress_bar=self.progress_bar)
+        else:
+            self._display_msg(inline_msg="Encoding not required.",
+                              progress_bar=self.progress_bar)
+        
+        # List of columns after one hot
+        col_af_ohe = self.data.columns
+        
+        # List of excluded columns from outlier processing and scaling
+        self.excluded_cols= self._extract_list(col_af_ohe, col_bf_ohe)
+        
+        end_time = time.time()
+        self._display_msg(msg="Time taken to encode the columns: {:.2f} sec".format( end_time - start_time),
+                          progress_bar=self.progress_bar,
+                          show_data=True)
+
+    def _custom_categorical_encoding(self):
+        """
+        DESCRIPTION:
+            Function to perform specific encoding on the categorical columns based on user input.
+            if validation fails, default encoding is getting performed on all remaining categorical columns.
+        """
+        self._display_msg(msg="\nStarting Customized Categorical Feature Encoding ...",
+                          progress_bar=self.progress_bar)
+        cat_end_input = self.custom_data.get("CategoricalEncodingIndicator", False)
+        # Checking user input for categorical encoding
+        if cat_end_input:
+            # Storing custom categorical encoding indicator in data transform dictionary
+            self.data_transform_dict["custom_categorical_encoding_ind"] = True
+            # Fetching user input list for performing 
+            encoding_list = self.custom_data.get("CategoricalEncodingParam", None)
+            if encoding_list:
+                onehot_encode_ind = encoding_list.get("OneHotEncodingIndicator", False)
+                ordinal_encode_ind = encoding_list.get("OrdinalEncodingIndicator", False)
+                target_encode_ind = encoding_list.get("TargetEncodingIndicator", False)
+                # Checking if any of categorical encoding technique indicator 
+                if not any([onehot_encode_ind, ordinal_encode_ind, target_encode_ind]):  
+                    self._display_msg(inline_msg="No information provided for any type of customized categorical encoding techniques. AutoML will proceed with default encoding technique.",
+                                      progress_bar=self.progress_bar)
+                else:
+                    if onehot_encode_ind:
+                        unique_count  = []
+                        ohe_list = encoding_list.get("OneHotEncodingList", None)
+                        # Checking for empty list
+                        if not ohe_list:
+                            self._display_msg(inline_msg="No information provided for customized one hot encoding technique.",
+                                              progress_bar=self.progress_bar)
+                        else:
+                            # Checking for column present in dataset or not
+                            _Validators._validate_dataframe_has_argument_columns(ohe_list, "OneHotEncodingList", self.data, "df")
+
+                            # Keeping track for existing columns before apply one hot encoding
+                            col_bf_ohe = self.data.columns
+                            # Detecting categorical columns with their unique counts
+                            for col in ohe_list:
+                                unique_count.append(self.data.drop_duplicate(col).size)
+                            # Performing one hot encoding   
+                            self._one_hot_encoding(ohe_list, unique_count)
+                            # Keeping track for new columns after apply one hot encoding
+                            col_af_ohe = self.data.columns
+                            # Fetching list of columns on which outlier processing should not be applied
+                            self.excluded_cols.extend(self._extract_list(col_af_ohe, col_bf_ohe))
+                            
+                            self._display_msg(msg="Updated dataset sample after performing one hot encoding:",
+                                              data=self.data,
+                                              progress_bar=self.progress_bar)
+                            
+                    if ordinal_encode_ind:
+                        ord_list = encoding_list.get("OrdinalEncodingList", None)
+                        # Checking for empty list
+                        if not ord_list:
+                            self._display_msg(inline_msg="No information provided for customized ordinal encoding technique.",
+                                              progress_bar=self.progress_bar)
+                        else:
+                            # Checking for column present in dataset or not
+                            _Validators._validate_dataframe_has_argument_columns(ord_list, "OrdinalEncodingList", self.data, "df")
+                            
+                            # Performing ordinal encoding
+                            self._ordinal_encoding(ord_list)
+                            self._display_msg(msg="Updated dataset sample after performing ordinal encoding:",
+                                              data=self.data,
+                                              progress_bar=self.progress_bar)
+
+                    if target_encode_ind:
+                        tar_list = encoding_list.get("TargetEncodingList", None)
+                        if not tar_list:
+                            self._display_msg(inline_msg="No information provided for customized target encoding technique.",
+                                              progress_bar=self.progress_bar)
+                        else:    
+                            # Performing target encoding
+                            self._target_encoding(tar_list)
+                            self._display_msg(msg="Updated dataset sample after performing target encoding:",
+                                              data=self.data,
+                                              progress_bar=self.progress_bar)
+            else:
+                self._display_msg(inline_msg="No input provided for performing customized categorical encoding. AutoML will proceed with default encoding technique.",
+                                  progress_bar=self.progress_bar)               
+        else:
+            self._display_msg(inline_msg="AutoML will proceed with default encoding technique.",
+                              progress_bar=self.progress_bar)
+            
+        # Performing default encoding on remaining categorical columns   
+        self._encoding_categorical_columns()
+
+    def _numapply_transformation(self, target_col, transform_val): 
+        """
+        DESCRIPTION:
+            Function to perform different numerical transformations using NumApply on numerical features based on user input.
+                    
+        """
+        # Fetching columns for accumulation
+        accumulate_columns = self._extract_list(self.data.columns, [target_col])
+        apply_method = transform_val["apply_method"]
+        # Adding fit parameters for performing transformation
+        fit_params={
+            "data": self.data,
+            "target_columns" : target_col,
+            "apply_method" : apply_method,
+            "inplace" : True,
+            "persist" :True,
+            "accumulate" : accumulate_columns
+        }
+        # Adding addition details for fit parameters in case of SIGMOID transformation
+        if apply_method == "sigmoid":
+            sigmoid_style=transform_val["sigmoid_style"]
+            fit_params = {**fit_params, "sigmoid_style" : sigmoid_style}
+        # Performing transformation on target columns
+        return NumApply(**fit_params).result
+  
+    def _numerical_transformation(self, target_columns, num_transform_data):
+        """
+        DESCRIPTION:
+            Function to perform different numerical transformations using Fit and Transform on numerical features based on user input.
+                    
+        """
+        # Adding fit parameters for transformation
+        fit_params={
+            "data" : self.data,
+            "object" : num_transform_data,
+            "object_order_column" : "TargetColumn"
+        }
+        # Peforming fit with all arguments.
+        num_fit_obj = Fit(**fit_params)
+        # Fetching all numerical columns
+        numerical_columns = [col for col, d_type in self.data._column_names_and_types if d_type in ["int","float"]]
+        # Extracting id columns where transformation should not affect numerical columns
+        id_columns = self._extract_list(numerical_columns,target_columns)
+        # Storing fit object and id column list for numerical transformation in data transform dictionary
+        self.data_transform_dict['custom_numerical_transformation_fit_object'] = num_fit_obj.result
+        self.data_transform_dict['custom_numerical_transformation_id_columns'] = id_columns
+        # Adding transform parameters for transformation
+        transform_params={
+            "data" : self.data,
+            "object" : num_fit_obj.result,
+            "id_columns" : id_columns,
+            "persist" :True
+        }
+        # Peforming transformation on target columns
+        self.data = Transform(**transform_params).result  
+        self._display_msg(msg="Updated dataset sample after applying numerical transformation:",
+                          data=self.data,
+                          progress_bar=self.progress_bar)
+    
+    def _mathematical_transformation(self):
+        """
+        DESCRIPTION:
+            Function to perform different mathematical transformations (i.e., log, pow,
+            exp, sininv, sigmoid) on numerical features based on user input.
+        """
+        self._display_msg(msg="\nStarting customized mathematical transformation ...",
+                          progress_bar=self.progress_bar,
+                          show_data=True)
+        
+        mat_transform_input = self.custom_data.get("MathameticalTransformationIndicator", False)
+        # Checking user input for mathematical transformations
+        if mat_transform_input:
+            # Extracting list required for mathematical transformations
+            mat_transform_list = self.custom_data.get("MathameticalTransformationParam", None)
+            if mat_transform_list:
+                # Checking for column present in dataset or not
+                _Validators._validate_dataframe_has_argument_columns(list(mat_transform_list.keys()), 
+                                                                     "MathameticalTransformationParam", self.data, "df")
+
+                # List of storing target columns and mathematical transformation information
+                transform_data=[]
+                target_columns=[]
+                # Storing custom mathematical transformation indicator in data transform dictionary
+                self.data_transform_dict['custom_mathematical_transformation_ind'] = True
+                # Storing custom numapply transformation parameters in data transform dictionary
+                self.data_transform_dict['custom_numapply_transformation_param'] = {}
+
+                for col, transform_val in mat_transform_list.items():
+                    apply_method=transform_val["apply_method"]
+                    if apply_method in (["sininv","sigmoid"]):
+                        # Applying numapply transformation
+                        self.data = self._numapply_transformation(col,transform_val)
+                        self._display_msg(msg="Updated dataset sample after applying numapply transformation:",
+                                          data=self.data,
+                                          progress_bar=self.progress_bar)
+                        # Updating parameter details for each column
+                        self.data_transform_dict['custom_numapply_transformation_param'].update({col:transform_val})
+                    else:
+                        # Handling specific scenarios for log and pow transformation
+                        parameters=""
+                        if apply_method == "log":
+                            base = transform_val["base"]
+                            parameters = json.dumps({"base":base})
+                        elif apply_method == "pow":
+                            exponent = transform_val["exponent"]
+                            parameters = json.dumps({"exponent":exponent})
+                        target_columns.append(col)
+                        transform_data.append({"TargetColumn":col, "DefaultValue":1, "Transformation":apply_method, "Parameters":parameters})
+                # Checking for transformation data
+                if len(transform_data):
+                    # Coverting into pandas and then teradata dataframe for performing further opration
+                    transform_data = pd.DataFrame(transform_data, columns=["TargetColumn", "DefaultValue", "Transformation", "Parameters"]) 
+                    self._display_msg(msg="Numerical transformation information :-",
+                                      data=transform_data,
+                                      progress_bar=self.progress_bar)
+                    copy_to_sql(df=transform_data, table_name="automl_num_transform_data", temporary=True)
+                    num_transform_data = DataFrame.from_table("automl_num_transform_data") 
+                    # Applying transformation using Fit/Transform functions
+                    self._numerical_transformation(target_columns, num_transform_data)
+                    # Storing custom numerical transformation parameters and column list in data transform dictionary
+                    self.data_transform_dict['custom_numerical_transformation_col'] = target_columns
+                    self.data_transform_dict['custom_numerical_transformation_params'] = num_transform_data
+            else:
+                self._display_msg(inline_msg="No input provided for performing customized mathematical transformation.",
+                                  progress_bar=self.progress_bar)
+        else:
+            self._display_msg(inline_msg="Skipping customized mathematical transformation.",
+                              progress_bar=self.progress_bar)
+
+    def _non_linear_transformation(self):
+        """
+        DESCRIPTION:
+            Function to perform customized non-linear transformation on numerical features based on user input.
+            
+        """
+        self._display_msg(msg="\nStarting customized non-linear transformation ...",
+                          progress_bar=self.progress_bar,
+                          show_data=True)
+        nl_transform_input = self.custom_data.get("NonLinearTransformationIndicator", False)
+        # Checking user input for non-linear transformation
+        if nl_transform_input:
+            nl_transform_list = self.custom_data.get("NonLinearTransformationParam", None)
+            # Extracting list required for non-linear transformation
+            if nl_transform_list:
+                total_combination = len(nl_transform_list)
+                # Generating all possible combination names
+                possible_combination = ["Combination_"+str(counter) for counter in range(1,total_combination+1)]
+                self._display_msg(msg="Possible combination :",
+                                  col_lst=possible_combination,
+                                  progress_bar=self.progress_bar)
+                # Storing custom non-linear transformation indicator in data transform dictionary
+                self.data_transform_dict['custom_non_linear_transformation_ind'] = True
+                # Storing custom non-linear transformation fit object in data transform dictionary
+                self.data_transform_dict['custom_non_linear_transformation_fit_object'] = {}
+                # print("Possible combination :",possible_combination)
+                # Performing transformation for each combination
+                for comb, transform_val in nl_transform_list.items():
+                    if comb in possible_combination:
+                        target_columns = transform_val["target_columns"]
+                        # Checking for column present in dataset or not
+                        _Validators._validate_dataframe_has_argument_columns(target_columns, 
+                                                                             "target_columns", self.data, "df")
+
+                        formula = transform_val["formula"]
+                        result_column = transform_val["result_column"]
+                        # Adding fit params for transformation
+                        fit_param = {
+                            "data" : self.data,
+                            "target_columns" : target_columns,
+                            "formula" : formula,
+                            "result_column" : result_column
+                        }
+                        # Performing fit on dataset
+                        fit_obj = NonLinearCombineFit(**fit_param)
+                        # Updating it for each non-linear combination
+                        self.data_transform_dict['custom_non_linear_transformation_fit_object'].update({comb:fit_obj})
+                        # Adding transform params for transformation   
+                        transform_params = {
+                            "data" : self.data,
+                            "object" : fit_obj,
+                            "accumulate" : self.data.columns,
+                            "persist" : True
+                        }
+                        self.data = NonLinearCombineTransform(**transform_params).result
+                    else:
+                        self._display_msg(inline_msg="Combinations are not as per expectation.",
+                                          progress_bar=self.progress_bar)
+                self._display_msg(msg="Updated dataset sample after performing non-liner transformation:",
+                                  data=self.data,
+                                  progress_bar=self.progress_bar)                  
+            else:
+                self._display_msg(inline_msg="No information provided for performing customized non-linear transformation.",
+                                  progress_bar=self.progress_bar)
+        else:
+            self._display_msg(inline_msg="Skipping customized non-linear transformation.",
+                              progress_bar=self.progress_bar)   
+            
+    def _anti_select_columns(self):
+        """
+        DESCRIPTION:
+            Function to remove specific features from dataset based on user input.
+                    
+        """
+        self._display_msg(msg="\nStarting customized anti-select columns ...",
+                          progress_bar=self.progress_bar,
+                          show_data=True)
+        anti_select_input = self.custom_data.get("AntiselectIndicator", False) 
+        # Checking user input for anti-select columns    
+        if anti_select_input:
+            # Extracting list required for anti-select columns
+            anti_select_list = self.custom_data.get("AntiselectParam", None)
+            if(anti_select_list):
+                if all(item in self.data.columns for item in anti_select_list):
+                    # Storing custom anti-select columns indicator and column list in data transform dictionary
+                    self.data_transform_dict['custom_anti_select_columns_ind'] = True
+                    self.data_transform_dict['custom_anti_select_columns'] = anti_select_list
+                    fit_params = {
+                        "data" : self.data,
+                        "exclude" : anti_select_list
+                    }
+                    # Performing transformation for given user input
+                    self.data = Antiselect(**fit_params).result
+                    self._display_msg(msg="Updated dataset sample after performing anti-select columns:",
+                                      data=self.data,
+                                      progress_bar=self.progress_bar)
+                else:
+                    self._display_msg(msg="Columns provided in list are not present in dataset:",
+                                      col_lst=anti_select_list,
+                                      progress_bar=self.progress_bar)         
+            else:
+                self._display_msg(inline_msg="No information provided for performing anti-select columns operation.",
+                                  progress_bar=self.progress_bar)
+        else:
+            self._display_msg(inline_msg="Skipping customized anti-select columns.",
+                              progress_bar=self.progress_bar)