numpyUtilsUpdated 0.0.2__tar.gz → 0.0.7__tar.gz

numpyutilsupdated-0.0.7/PKG-INFO

@@ -0,0 +1,34 @@
+Metadata-Version: 2.4
+Name: numpyUtilsUpdated
+Version: 0.0.7
+Summary: A collection of numpy utilities for data analysis and manipulation.
+Project-URL: Homepage, https://github.com/nani-here/numpyUtilsUpdated
+Project-URL: Issues, https://github.com/nani-here/numpyUtilsUpdated/issues
+Author-email: Nani bolthe! <neku.enduku2005@gmail.com>
+License-File: LICENSE
+Classifier: License :: OSI Approved :: GNU General Public License v3 (GPLv3)
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Requires-Python: >=3.0
+Description-Content-Type: text/markdown
+
+# numpyUtilsUpdated
+
+A comprehensive collection of utility functions designed to enhance your NumPy workflows. This module provides a range of tools to simplify common NumPy operations, improve code readability, and boost your productivity when working with numerical data in Python.
+
+This module provides questions and Python codes for ML lab sets.
+
+## Install Module
+
+```bash
+pip install numpyUtilsUpdated
+
+then import like this:
+
+```bash
+import numpyUtilsUpdated as nup
+
+the module will give further instructions
+
+```bash
+pip install numpy pandas scikit-learn statsmodels matplotlib seaborn scipy

numpyutilsupdated-0.0.7/README.md

@@ -0,0 +1,20 @@
+# numpyUtilsUpdated
+
+A comprehensive collection of utility functions designed to enhance your NumPy workflows. This module provides a range of tools to simplify common NumPy operations, improve code readability, and boost your productivity when working with numerical data in Python.
+
+This module provides questions and Python codes for ML lab sets.
+
+## Install Module
+
+```bash
+pip install numpyUtilsUpdated
+
+then import like this:
+
+```bash
+import numpyUtilsUpdated as nup
+
+the module will give further instructions
+
+```bash
+pip install numpy pandas scikit-learn statsmodels matplotlib seaborn scipy

{numpyutilsupdated-0.0.2 → numpyutilsupdated-0.0.7}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "numpyUtilsUpdated"
-version = "0.0.2"
+version = "0.0.7"
 authors = [
   { name="Nani bolthe!", email="neku.enduku2005@gmail.com" },
 ]

numpyutilsupdated-0.0.7/src/numpyUtilsUpdated/__init__.py

@@ -0,0 +1,102 @@
+print("[numpyUtilsUpdated] : Welcome to LM Codes")
+print("!pip install numpy pandas scikit-learn statsmodels matplotlib seaborn scipy --user")
+
+print("""
+Use the function:
+      
+import numpyUtilsUpdated as nup
+nup.getSetDetails(set_no)
+
+Example:
+getSetDetails(1)
+
+This will print the question for that set.
+
+Then to get the code use:
+Replace set1 with your set no, [set1 - set18] 
+import numpyUtilsUpdated.set1 as set
+print(set.code)
+""")
+
+SET_DETAILS = {
+    "set1": {
+        "question": "Develop an SLR model to predict MBA salary using Grade 10 percentage from MBA salary.csv. Diagnose the regression model and perform residual analysis using P-P plot."
+    },
+
+    "set2": {
+        "question": "Develop a simple linear regression model between Corruption Perception Index (Y) and Gini Index (X) using country.csv. Diagnose the regression model and perform residual analysis using P-P plot."
+    },
+
+    "set3": {
+        "question": "Develop an SLR model using MBA salary.csv and detect outliers using Z-score and Cook’s distance. Make prediction and measure accuracy."
+    },
+
+    "set4": {
+        "question": "Using IPL dataset, build an MLR model, show the summary, and identify features with multicollinearity."
+    },
+
+    "set5": {
+        "question": "Using IPL dataset, build an MLR model, detect multicollinearity, rebuild the model after removing it, and perform residual analysis using P-P plot."
+    },
+
+    "set6": {
+        "question": "Using country.csv dataset, build an SLR model between Corruption Index and Gini Index, detect outliers using Z-score and Cook’s distance, and evaluate prediction accuracy."
+    },
+
+    "set7": {
+        "question": "Using GermanCredit.csv dataset, build a logistic regression model to predict credit risk, identify significant features, rebuild the model, compute confusion matrix, precision, recall, ROC and AUC."
+    },
+
+    "set8": {
+        "question": "Using GermanCredit.csv dataset, build logistic regression, compute Youden’s index for cut-offs from 0.1 to 0.5, find optimal cut-off, build confusion matrix, and compute ROC and AUC."
+    },
+
+    "set9": {
+        "question": "Demonstrate Gain and Lift charts using bank.csv dataset."
+    },
+
+    "set10": {
+        "question": "Using the tennis dataset, build a logistic regression model to predict PLAY, identify significant features, analyze coefficients, and compute confusion matrix with precision and recall."
+    },
+
+    "set11": {
+        "question": "Construct the decision tree using Gini impurity for the given training dataset."
+    },
+
+    "set12": {
+        "question": "Demonstrate Gradient Descent Algorithm for Linear Regression using Advertising.csv dataset."
+    },
+
+    "set13": {
+        "question": "Build logistic regression models on bank.csv dataset for both imbalanced and balanced data, evaluate using 5-fold cross-validation and ROC AUC score."
+    },
+
+    "set14": {
+        "question": "Demonstrate the KNN algorithm using a suitable dataset."
+    },
+
+    "set15": {
+        "question": "Using Income Data.csv dataset, perform K-Means clustering, draw scatter plot of age vs income, normalize features, plot clusters, and interpret cluster centers."
+    },
+
+    "set16": {
+        "question": "Using customerspends.csv dataset, perform K-Means clustering, visualize clusters, normalize features, use dendrogram and elbow method, and print cluster centers."
+    },
+
+    "set17": {
+        "question": "Using Income Data.csv dataset, identify clusters, normalize features, apply elbow method, and print records and cluster centers."
+    },
+
+    "set18": {
+        "question": "Perform product segmentation using K-Means clustering on customerspends.csv dataset, visualize clusters, verify with dendrogram and elbow method, and print cluster centers."
+    }
+}
+
+def getSetDetails(set_no):
+    key = f"set{set_no}"
+
+    if key in SET_DETAILS:
+        print(f"Set {set_no} Question:\n")
+        print(SET_DETAILS[key]["question"])
+    else:
+        print("Invalid set number")

numpyutilsupdated-0.0.7/src/numpyUtilsUpdated/set1.py

@@ -0,0 +1,94 @@
+code="""
+# ============================================================
+# COMMON IMPORTS — Run this cell FIRST before any question
+# ============================================================
+import pandas as pd
+import numpy as np
+import statsmodels.api as sm
+import matplotlib.pyplot as plt
+import seaborn as sns
+from scipy.stats import zscore
+from sklearn.model_selection import train_test_split, cross_val_score, GridSearchCV
+from sklearn.linear_model import LogisticRegression
+from sklearn import metrics
+from sklearn.neighbors import KNeighborsClassifier
+from sklearn.cluster import KMeans
+from sklearn.preprocessing import StandardScaler
+from sklearn.utils import resample, shuffle
+from statsmodels.stats.outliers_influence import variance_inflation_factor
+from statsmodels.graphics.regressionplots import influence_plot
+from scipy.cluster.hierarchy import dendrogram, linkage
+
+print('All imports successful!')
+
+# ============================================================
+# QUESTION 1 — SLR on MBA Salary Dataset
+# ============================================================
+
+# ---------- Step 1: Load Dataset ----------
+mba_df = pd.read_csv('MBA_salary.csv')
+print('First 5 rows:')
+print(mba_df.head())
+print('\nDataset Info:')
+print(mba_df.info())
+
+# ---------- Step 2 (i): Build SLR Model ----------
+# X = Grade 10 percentage (independent variable)
+# y = Salary (dependent variable)
+X = sm.add_constant(mba_df['percentage in Grade 10'])   # adds intercept column
+y = mba_df['salary']
+
+# Split: 80% train, 20% test
+X_train, X_test, y_train, y_test = train_test_split(
+    X, y, train_size=0.8, random_state=42)
+
+# Fit OLS (Ordinary Least Squares) model
+mba_lm = sm.OLS(y_train, X_train).fit()
+print('\n===== MODEL SUMMARY (i) =====')
+print(mba_lm.summary2())
+
+# ---------- Step 3 (ii): Diagnose the Model — Homoscedasticity ----------
+mba_resid = mba_lm.resid   # residuals = actual − predicted
+
+def get_std_values(vals):
+    return (vals - vals.mean()) / vals.std()
+
+plt.figure(figsize=(8, 5))
+plt.scatter(
+    get_std_values(mba_lm.fittedvalues),
+    get_std_values(mba_resid)
+)
+plt.axhline(y=0, color='red', linestyle='--')
+plt.title('(ii) Residual Plot — Homoscedasticity Check')
+plt.xlabel('Standardized Predicted Values')
+plt.ylabel('Standardized Residuals')
+plt.show()
+
+# ---------- Step 4 (iii): P-P Plot — Residual Normality ----------
+probplot = sm.ProbPlot(mba_resid)
+plt.figure(figsize=(8, 5))
+probplot.ppplot(line='45')
+plt.title('(iii) Normal P-P Plot of Regression Standardized Residuals')
+plt.show()
+
+print('''
+EXPLANATION:
+(i)  sm.add_constant() adds an intercept (β₀) column.
+     sm.OLS().fit()    trains the linear model on training data.
+     summary2()        shows R², coefficients, p-values, AIC, BIC.
+     • R²: % of salary variation explained by Grade 10 %.
+     • p-value < 0.05 → feature is statistically significant.
+
+(ii) Residual Plot checks Homoscedasticity:
+     → Residuals randomly scattered around 0 = model is valid.
+     → Pattern or funnel shape = Heteroscedasticity (problem).
+
+(iii) P-P Plot checks if residuals are Normally distributed:
+     → Points close to 45° line = residuals are normal = model is valid.
+     → Points far from line = residuals are NOT normal.
+''')
+
+END
+"""
+def getCode():
+    print(code)

numpyutilsupdated-0.0.7/src/numpyUtilsUpdated/set10.py

@@ -0,0 +1,135 @@
+code="""
+# ============================================================
+# COMMON IMPORTS — Run this cell FIRST before any question
+# ============================================================
+import pandas as pd
+import numpy as np
+import statsmodels.api as sm
+import matplotlib.pyplot as plt
+import seaborn as sns
+from scipy.stats import zscore
+from sklearn.model_selection import train_test_split, cross_val_score, GridSearchCV
+from sklearn.linear_model import LogisticRegression
+from sklearn import metrics
+from sklearn.neighbors import KNeighborsClassifier
+from sklearn.cluster import KMeans
+from sklearn.preprocessing import StandardScaler
+from sklearn.utils import resample, shuffle
+from statsmodels.stats.outliers_influence import variance_inflation_factor
+from statsmodels.graphics.regressionplots import influence_plot
+from scipy.cluster.hierarchy import dendrogram, linkage
+
+print('All imports successful!')
+
+# ============================================================
+# QUESTION 10 — Logistic Regression on Tennis Dataset
+# ============================================================
+
+# ---------- Step 1: Create Dataset from Question ----------
+tennis_data = {
+    'DAY'     : [f'Day{i}' for i in range(1, 15)],
+    'OUTLOOK' : ['Sunny','Sunny','Overcast','Rain','Rain','Rain',
+                 'Overcast','Sunny','Sunny','Rain','Sunny','Overcast','Overcast','Rain'],
+    'TEMP'    : ['Hot','Hot','Hot','Mild','Cool','Cool','Cool',
+                 'Mild','Cool','Mild','Mild','Mild','Hot','Mild'],
+    'HUMIDITY': ['High','High','High','High','Normal','Normal','Normal',
+                 'High','Normal','Normal','Normal','High','Normal','High'],
+    'WIND'    : ['Weak','Strong','Weak','Weak','Weak','Strong','Strong',
+                 'Weak','Weak','Weak','Strong','Strong','Weak','Strong'],
+    'PLAY'    : ['NO','NO','YES','YES','YES','NO','YES',
+                 'NO','YES','YES','YES','YES','YES','NO']
+}
+tennis_df = pd.DataFrame(tennis_data)
+print('Tennis Dataset:')
+print(tennis_df.to_string(index=False))
+
+# ---------- Step 2 (i): Build Logistic Model ----------
+tennis_enc = pd.get_dummies(
+    tennis_df[['OUTLOOK','TEMP','HUMIDITY','WIND']],
+    drop_first=True, dtype=int)
+y_tennis = (tennis_df['PLAY'] == 'YES').astype(int)
+
+# Use sklearn since n=14 is very small (statsmodels may not converge)
+tennis_clf = LogisticRegression(max_iter=10000, solver='lbfgs')
+tennis_clf.fit(tennis_enc, y_tennis)
+
+# Show all features and coefficients
+coeff_tennis = pd.DataFrame({
+    'feature'    : tennis_enc.columns,
+    'coefficient': tennis_clf.coef_[0]
+}).sort_values('coefficient', ascending=False)
+
+print('\n===== (i) All Features and Coefficients =====')
+print(coeff_tennis.to_string(index=False))
+
+# ---------- Step 3 (ii): Significant Features ----------
+# For small datasets use statsmodels for p-values
+X_tennis_sm = sm.add_constant(tennis_enc)
+try:
+    logit_sm = sm.Logit(y_tennis, X_tennis_sm).fit(maxiter=500, disp=False)
+    sig_tennis = [v for v, p in logit_sm.pvalues.items() if p <= 0.05]
+    print(f'\n(ii) Significant features (p ≤ 0.05): {sig_tennis}')
+except:
+    # Fallback: use features with largest absolute coefficients
+    sig_tennis = list(coeff_tennis.nlargest(3, 'coefficient')['feature']) + \
+                 list(coeff_tennis.nsmallest(2, 'coefficient')['feature'])
+    print(f'\n(ii) Top features by coefficient magnitude: {sig_tennis}')
+
+# Build new model with significant features
+tennis_clf2 = LogisticRegression(max_iter=10000)
+sig_tennis_cols = [c for c in sig_tennis if c in tennis_enc.columns and c != 'const']
+if sig_tennis_cols:
+    tennis_clf2.fit(tennis_enc[sig_tennis_cols], y_tennis)
+    print(f'New model features: {sig_tennis_cols}')
+else:
+    tennis_clf2 = tennis_clf
+    sig_tennis_cols = list(tennis_enc.columns)
+
+# ---------- Step 4 (iii): Positive/Negative Effects ----------
+print('\n===== (iii) Effect on Probability of PLAY =====')
+for _, row in coeff_tennis.iterrows():
+    effect = 'POSITIVE (+) — increases PLAY probability' if row['coefficient'] > 0 \
+             else 'NEGATIVE (−) — decreases PLAY probability'
+    print(f"  {row['feature']:30s}: coeff={row['coefficient']:+.4f}  →  {effect}")
+
+# ---------- Step 5 (iv): Confusion Matrix at 0.5 Cut-off ----------
+pred_tennis = tennis_clf.predict(tennis_enc)   # uses default 0.5 threshold
+
+cm_tennis = metrics.confusion_matrix(y_tennis, pred_tennis)
+plt.figure(figsize=(6, 5))
+sns.heatmap(cm_tennis, annot=True, fmt='.0f',
+            xticklabels=['NO', 'YES'],
+            yticklabels=['NO', 'YES'],
+            cmap='Blues')
+plt.xlabel('Predicted')
+plt.ylabel('Actual')
+plt.title('(iv) Confusion Matrix — Tennis (cut-off = 0.5)')
+plt.show()
+
+print('\n(iv) Classification Report (PLAY = YES):')
+print(metrics.classification_report(y_tennis, pred_tennis,
+      target_names=['NO', 'YES']))
+
+print('''
+EXPLANATION:
+(i)  get_dummies() converts OUTLOOK (Sunny/Overcast/Rain) → binary columns.
+     OUTLOOK_Overcast=1, OUTLOOK_Sunny=1, TEMP_Hot=1, WIND_Weak=1 etc.
+     drop_first=True: removes one level per feature to avoid multicollinearity.
+
+(ii) Only 14 training samples → statsmodels may struggle to converge.
+     Significant features are those with the strongest influence.
+
+(iii) Positive coefficient:
+      e.g., HUMIDITY_Normal = positive → Normal humidity increases chance of PLAY.
+      Negative coefficient:
+      e.g., WIND_Strong = negative → Strong wind decreases chance of PLAY.
+
+(iv) Precision for PLAY=YES:
+     Of all days we predicted play, how many actually played.
+     Recall for PLAY=YES:
+     Of all actual play days, how many did we correctly predict.
+''')
+
+END
+
+"""

numpyutilsupdated-0.0.7/src/numpyUtilsUpdated/set11.py

@@ -0,0 +1,82 @@
+code = """
+# =====================================
+# Decision Tree using Gini Impurity
+# =====================================
+
+import pandas as pd
+import matplotlib.pyplot as plt
+import seaborn as sns
+from sklearn.tree import DecisionTreeClassifier, plot_tree
+from sklearn.metrics import confusion_matrix, classification_report, accuracy_score
+
+# -------------------------------------
+# DATASET
+# -------------------------------------
+data = {
+'Weather':['Sunny','Cloudy','Sunny','Cloudy','Rainy','Rainy','Rainy','Sunny','Cloudy','Rainy'],
+'Temperature':['Hot','Hot','Mild','Mild','Mild','Cool','Mild','Hot','Hot','Mild'],
+'Humidity':['High','High','Normal','High','High','Normal','High','High','Normal','High'],
+'Wind':['Weak','Weak','Strong','Strong','Strong','Strong','Weak','Strong','Weak','Strong'],
+'Play':['No','Yes','Yes','Yes','No','No','Yes','No','Yes','No']
+}
+
+df = pd.DataFrame(data)
+print("Dataset:\n",df)
+
+# -------------------------------------
+# ENCODE DATA
+# -------------------------------------
+df = pd.get_dummies(df)
+
+X = df.drop('Play_Yes',axis=1)
+y = df['Play_Yes']
+
+# -------------------------------------
+# TRAIN MODEL
+# -------------------------------------
+model = DecisionTreeClassifier(criterion='gini',max_depth=3)
+model.fit(X,y)
+
+# -------------------------------------
+# TREE GRAPH
+# -------------------------------------
+plt.figure(figsize=(12,6))
+plot_tree(model,feature_names=X.columns,class_names=['No','Yes'],filled=True)
+plt.title("Decision Tree (Gini Impurity)")
+plt.show()
+
+# -------------------------------------
+# PREDICTIONS
+# -------------------------------------
+y_pred = model.predict(X)
+
+print("\nAccuracy:",accuracy_score(y,y_pred))
+print("\nClassification Report:\n",classification_report(y,y_pred))
+
+# -------------------------------------
+# CONFUSION MATRIX (DIAGNOSTIC)
+# -------------------------------------
+cm = confusion_matrix(y,y_pred)
+
+plt.figure(figsize=(5,4))
+sns.heatmap(cm,annot=True,fmt='d',cmap='Blues',
+            xticklabels=['No','Yes'],
+            yticklabels=['No','Yes'])
+plt.xlabel("Predicted")
+plt.ylabel("Actual")
+plt.title("Confusion Matrix")
+plt.show()
+
+# -------------------------------------
+# FEATURE IMPORTANCE GRAPH
+# -------------------------------------
+importance = pd.Series(model.feature_importances_,index=X.columns)
+
+importance.sort_values().plot(kind='barh',figsize=(8,5))
+plt.title("Feature Importance (Gini Reduction)")
+plt.xlabel("Importance")
+plt.show()
+
+END
+
+"""

numpyutilsupdated-0.0.7/src/numpyUtilsUpdated/set12.py

@@ -0,0 +1,135 @@
+code='''
+# ============================================================
+# COMMON IMPORTS — Run this cell FIRST before any question
+# ============================================================
+import pandas as pd
+import numpy as np
+import statsmodels.api as sm
+import matplotlib.pyplot as plt
+import seaborn as sns
+from scipy.stats import zscore
+from sklearn.model_selection import train_test_split, cross_val_score, GridSearchCV
+from sklearn.linear_model import LogisticRegression
+from sklearn import metrics
+from sklearn.neighbors import KNeighborsClassifier
+from sklearn.cluster import KMeans
+from sklearn.preprocessing import StandardScaler
+from sklearn.utils import resample, shuffle
+from statsmodels.stats.outliers_influence import variance_inflation_factor
+from statsmodels.graphics.regressionplots import influence_plot
+from scipy.cluster.hierarchy import dendrogram, linkage
+
+print('All imports successful!')
+
+# ============================================================
+# QUESTION 12 — Gradient Descent Algorithm for Linear Regression
+# ============================================================
+
+# ---------- Step 1: Load Dataset ----------
+sales_df = pd.read_csv('Advertising.csv')
+print('Dataset:')
+print(sales_df.head())
+print('Shape:', sales_df.shape)
+
+X_sales = sales_df[['TV', 'Radio', 'Newspaper']]
+y_sales = sales_df['sales']
+
+# ---------- Step 2: Standardize Features ----------
+y_std = np.array((y_sales - y_sales.mean()) / y_sales.std())
+X_std = X_sales.apply(
+    lambda col: (col - col.mean()) / col.std(), axis=0)
+
+# ---------- Step 3: Define Gradient Descent Functions ----------
+
+def initialize(dim):
+    """Random initialization of weights (w) and bias (b)"""
+    np.random.seed(42)
+    b = np.random.random()
+    w = np.random.rand(dim)
+    return b, w
+
+def predict_y(b, w, X):
+    """Forward pass: y_hat = b + X·w"""
+    return b + np.matmul(X, w)
+
+def get_cost(y, y_hat):
+    """Cost function: Mean Squared Error (MSE)"""
+    residuals = y - y_hat
+    return np.sum(np.matmul(residuals.T, residuals)) / len(residuals)
+
+def update_beta(X, y, y_hat, b0, w0, learning_rate):
+    """Gradient update step for bias and weights"""
+    db = (np.sum(y_hat - y) * 2) / len(y)           # gradient for bias
+    dw = (np.dot((y_hat - y), X) * 2) / len(y)      # gradient for weights
+    b1 = b0 - learning_rate * db                     # update bias
+    w1 = w0 - learning_rate * dw                     # update weights
+    return b1, w1
+
+def run_grad(X, y, alpha=0.01, num_iterations=100):
+    """Run full Gradient Descent for given iterations and learning rate"""
+    b, w        = initialize(X.shape[1])
+    iter_num    = 0
+    gd_iter_df  = pd.DataFrame(columns=['iterations', 'cost'])
+    result_idx  = 0
+
+    for iter_num in range(num_iterations):
+        y_hat     = predict_y(b, w, X)
+        this_cost = get_cost(y, y_hat)
+        prev_b, prev_w = b, w
+        b, w      = update_beta(X, y, y_hat, prev_b, prev_w, alpha)
+
+        if iter_num % 10 == 0:   # record every 10th iteration
+            gd_iter_df.loc[result_idx] = [iter_num, this_cost]
+            result_idx += 1
+
+    print(f'Final estimate of b & w: {round(b,5)}  {np.round(w,5)}')
+    return gd_iter_df, b, w
+
+# ---------- Step 4: Show Initial Parameters ----------
+b_init, w_init = initialize(3)
+print(f'\nInitial Bias    : {b_init:.4f}')
+print(f'Initial Weights : {w_init}')
+
+y_hat_init = predict_y(b_init, w_init, X_std.values)
+cost_init  = get_cost(y_std, y_hat_init)
+print(f'Initial Cost (MSE): {cost_init:.4f}')
+
+b_init, w_init = update_beta(
+    X_std.values, y_std, y_hat_init, b_init, w_init, 0.01)
+print(f'\nAfter first update → Bias: {b_init:.4f}, Weights: {np.round(w_init,4)}')
+
+# ---------- Step 5: Run Gradient Descent ----------
+print('\n===== Running Gradient Descent (alpha=0.01, 2000 iterations) =====')
+gd_df1, b1, w1 = run_grad(X_std.values, y_std, alpha=0.01, num_iterations=2000)
+print('\nCost per 10 iterations (first 40 rows):')
+print(gd_df1.head(40).to_string(index=False))
+
+# ---------- Step 6: Plot Cost vs Iterations ----------
+print('\n===== Running with alpha=0.001 for comparison =====')
+gd_df2, b2, w2 = run_grad(X_std.values, y_std, alpha=0.001, num_iterations=2000)
+
+plt.figure(figsize=(12, 5))
+plt.subplot(1, 2, 1)
+plt.plot(gd_df1['iterations'].astype(float),
+         gd_df1['cost'].astype(float), color='blue', label='alpha=0.01')
+plt.xlabel('No. of Iterations')
+plt.ylabel('Cost (MSE)')
+plt.title('Cost vs Iterations (alpha=0.01)')
+plt.legend()
+plt.grid(True, alpha=0.3)
+
+plt.subplot(1, 2, 2)
+plt.plot(gd_df2['iterations'].astype(float),
+         gd_df2['cost'].astype(float), color='orange', label='alpha=0.001')
+plt.xlabel('No. of Iterations')
+plt.ylabel('Cost (MSE)')
+plt.title('Cost vs Iterations (alpha=0.001)')
+plt.legend()
+plt.grid(True, alpha=0.3)
+
+plt.tight_layout()
+plt.show()
+
+END
+
+'''