teradataml 20.0.0.8__py3-none-any.whl

teradataml/data/jsons/uaf/17.20/TD_AUTOARIMA.json

@@ -0,0 +1,469 @@
+{
+        "FuncName": "TD_AUTOARIMA",
+        "FuncDescriptionShort": "TD_AUTOARIMA fits the best ARIMA model to univariate time series.",
+        "FuncDescriptionLong": [
+                "The function searches the possible models within the order constrains in the function parameters.",
+                "returns the best ARIMA model based on the criterion provided by the INFOR_CRITERIA parameter.",       
+                "The TD_AUTOARIMA function creates a six-layered ART table."
+        ],
+        "FunctionVersion": "...",
+        "FunctionCategory": "Model Preparation and Parameter Estimation",
+        "JSONVersion": "1",
+        "FuncRName": "td_Auto_Arima",
+        "MaxInputFiles": 1,
+        "Input": [
+                {
+                        "Type": "SERIES",
+                        "Description": [
+                                "Time series whose value can be REAL."
+                        ],
+                        "LangName": "data or object or newdata or ... --> Langauges team can work with UAF team to come up with this field",
+                        "Optional": false
+                }
+        ],
+        "Output": [
+                {
+                        "Type": "ART",
+                        "PrimaryLayer": true,
+                        "LayerName": "ARTPRIMARY",
+                        "ResultTableColumnTypes": [
+								"<varies>",
+								"big_integer",
+                                "VARCHAR",
+                                "float",
+                                "float",
+                                "float",
+								"float"
+                        ],
+                        "Description": [
+                                "The primary result containing the selected best model’s coefficients.",
+                                "The primary result set is accessed using a SELECT statement."
+                        ],
+                        "LangName": "data or object or newdata or ... --> Langauges team can work with UAF team to come up with this field"
+                },           
+                {
+                        "Type": "ART",
+                        "PrimaryLayer": false,
+                        "LayerName": "ARTFITMETADATA",
+                        "ResultTableColumnTypes": [
+								"<varies>",
+								"big_integer",
+                                "integer",
+                                "integer",
+                                "float",
+                                "float",
+                                "float",
+                                "float",
+                                "float",
+                                "float",
+                                "float",
+                                "float",
+                                "float",
+                                "float",
+                                "float",
+                                "float",
+                                "float",
+                                "float",
+								"float",
+								"float",
+								"VARCHAR"
+                        ],
+                        "Description": [
+                                "The secondary result set containing “goodness of fit” metadata, is described in this subsection.   This result set is generated in response to the FUNC_PARAMS (FIT_METRICS(1) having a passed in flag value of '1'.   The result set can be retrieved by invoking the TD_EXTRACT_RESULTS function against the analytical result table containing the metadata. ",
+                                "There is one result row returned for each unique series acted upon. "
+                        ],
+                        "LangName": "data or object or newdata or ... --> Langauges team can work with UAF team to come up with this field"
+                },
+				{
+                        "Type": "ART",
+                        "PrimaryLayer": false,
+                        "LayerName": "ARTFITRESIDUALS",
+                        "ResultTableColumnTypes": [
+                                "<varies>",
+								"big_integer",
+								"float",
+								"float",
+								"float"
+                        ],
+                        "Description": [
+                                "The tertiary result set contains residuals from the fitting exercise.",
+                                "The tertiary is accessed using the TD_EXTRACT_RESULTS utility function."
+                        ],
+                        "LangName": "data or object or newdata or ... --> Langauges team can work with UAF team to come up with this field"
+                },
+				{
+                        "Type": "ART",
+                        "PrimaryLayer": false,
+                        "LayerName": "ARTMODEL",
+                        "ResultTableColumnTypes": [
+                                "<varies>",
+								"big_integer",
+								"VARBYTE"
+                        ],
+                        "Description": [
+                                "The quaternary result set contains the best model context, which is used during the forecasting process.",
+                                "The quaternary is accessed using the TD_EXTRACT_RESULTS utility function."
+                        ],
+                        "LangName": "data or object or newdata or ... --> Langauges team can work with UAF team to come up with this field"
+                },
+				{
+                        "Type": "ART",
+                        "PrimaryLayer": false,
+                        "LayerName": "ARTICANDORDER",
+                        "ResultTableColumnTypes": [
+                                "<varies>",
+								"big_integer",
+								"integer",
+                                "integer",
+                                "float",
+                                "float",
+                                "float",
+                                "float",
+								"float",
+								"VARCHAR"
+                        ],
+                        "Description": [
+                                "The quinary result set contains the information criteria such as AIC and SBIC, and the order of the best model.",
+                                "The quinary is accessed using the TD_EXTRACT_RESULTS utility function."
+                        ],
+                        "LangName": "data or object or newdata or ... --> Langauges team can work with UAF team to come up with this field"
+                },
+				{
+                        "Type": "ART",
+                        "PrimaryLayer": false,
+                        "LayerName": "ARTARMAROOTS",
+                        "ResultTableColumnTypes": [
+                                "<varies>",
+								"big_integer",
+								"VARCHAR",                        
+                                "float",
+                                "float",
+                                "VARCHAR"                             					
+                        ],
+                        "Description": [
+                                "The senary result set contains the roots information.",
+                                "The senary is accessed using the TD_EXTRACT_RESULTS utility function."
+                        ],
+                        "LangName": "data or object or newdata or ... --> Langauges team can work with UAF team to come up with this field"
+                }
+        ],
+        "IsPlottable": true,
+        "Params": [
+                {
+                        "Name": "MAX_PQ_NONSEASONAL",
+                        "Type": "list",
+                        "ListType": "integer",
+                        "ListSize": 2,
+                        "Optional": true,
+                        "LowerBound": 0,
+                        "LowerBoundType": "INCLUSIVE",
+						"DefaultValue": [5, 5],
+                        "Description": [
+                                "The (p,q) order of the maximum autoregression (AR) and moving average (MA) parameters."
+                        ]
+                },
+				{
+                        "Name": "MAX_PQ_SEASONAL",
+                        "Type": "list",
+                        "ListType": "integer",
+                        "ListSize": 2,
+                        "Optional": true,
+                        "LowerBound": 0,
+                        "LowerBoundType": "INCLUSIVE",
+						"DefaultValue": [2, 2],
+                        "Description": [
+                                "The (P,Q) order of the max seasonal AR and MA parameters."
+                        ]
+                },
+				{
+                        "Name": "START_PQ_NONSEASONAL",
+                        "Type": "list",
+                        "ListType": "integer",
+                        "ListSize": 2,
+                        "Optional": true,       
+						"LowerBound": 0,
+                        "LowerBoundType": "INCLUSIVE",						
+						"DefaultValue": [0, 0],
+                        "Description": [
+                                "The start value of (p,q). Only used when STEPWISE(1)."
+                        ]
+                },
+				{
+                        "Name": "START_PQ_SEASONAL",
+                        "Type": "list",
+                        "ListType": "integer",
+                        "ListSize": 2,
+                        "Optional": true,
+                        "LowerBound": 0,
+                        "LowerBoundType": "INCLUSIVE",
+						"DefaultValue": [0, 0],
+                        "Description": [
+                                "The start value of seasonal (P,Q). Only used when STEPWISE(1)."
+                        ]
+                },
+				{
+                        "Name" : "d",
+						"Type" : "integer",
+						"Optional" : true,                     
+						"DefaultValue": -1,
+                        "Description": [
+                                "The order of first-differencing. Default is -1 (auto search d)."
+                        ]
+                },
+				{
+                        "Name" : "Ds",
+						"Type" : "integer",
+						"Optional" : true,                     
+						"DefaultValue": -1,
+                        "Description": [
+                                "The order of seasonal-differencing. Default: -1 (auto search Ds))."
+                        ]
+                },
+				{
+                        "Name" : "MAX_d",
+						"Type" : "integer",
+						"Optional" : true,
+                        "LowerBound": 0,												
+                        "LowerBoundType": "INCLUSIVE",
+						"DefaultValue": 2,
+                        "Description": [
+                                "Maximum number of non-seasonal differences."
+                        ]
+                },
+				{
+                        "Name" : "MAX_Ds",
+						"Type" : "integer",
+						"Optional" : true,
+                        "LowerBound": 0,												
+                        "LowerBoundType": "INCLUSIVE",
+						"DefaultValue": 1,
+                        "Description": [
+                                "Maximum number of seasonal differences."
+                        ]
+                },
+				{
+                        "Name" : "PERIOD",
+						"Type" : "integer",
+						"Optional" : true,
+                        "LowerBound": 1,
+						"UpperBound": 12,
+                        "LowerBoundType": "INCLUSIVE",
+                        "UpperBoundType": "INCLUSIVE",
+						"DefaultValue": 1,
+                        "Description": [
+                                "The number of periods per season. For non-seasonal data, period is 1."
+                        ]
+                },
+				{
+                        "Name" : "STATIONARY",
+						"Type" : "integer",
+						"Optional" : true,
+                        "PermittedValues": [
+                               0,
+							   1
+                        ],
+						"DefaultValue": 0,
+                        "Description": [
+                                "If true, the the function restricts search to stationary models. A value of 1 means true. A value of 0 means false."
+                        ]
+                },
+				{
+                        "Name" : "SEASONAL",
+						"Type" : "integer",
+						"Optional" : true,
+                        "PermittedValues": [
+                               0,
+							   1
+                        ],
+						"DefaultValue": 1,
+                        "Description": [
+                                "If false, then the function restricts search to non-seasonal models. A value of 1 means true. A value of 0 means false. Default is 1, meaning search all models including seasonal."
+                        ]
+                },
+				{
+                        "Name" : "CONSTANT",
+						"Type" : "integer",
+						"Optional" : true,
+                        "PermittedValues": [
+                               0,
+							   1
+                        ],
+						"DefaultValue": 1,
+                        "Description": [
+                                "Indicator that TD_AUTOARIMA function include an intercept. A value of 1 means CONSTANT/intercept should be included. A value of 0 means CONSTANT/intercept should not be included."
+                        ]
+                },
+				{
+                        "Name" : "ALGORITHM",
+						"Type" : "string",
+						"Optional" : true,
+                        "PermittedValues": [
+                               "MLE", 
+							   "CSS_MLE", 
+							   "CSS"
+                        ],
+						"DefaultValue": "MLE",
+                        "Description": [
+                                "The approach used by TD_AUTOARIMA to estimate the coefficients."
+                        ]
+                },
+				{
+                        "Name" : "FIT_PERCENTAGE",
+						"Type" : "integer",
+						"Optional" : true,
+                        "PermittedValues": [
+                               
+                        ],
+						"DefaultValue": 100,
+                        "Description": [
+                                "Percentage of passed-in sample points used for the model fitting (parameter estimation). The default is 100, meaning 100%."
+                        ]
+                },
+				{
+                        "Name" : "INFOR_CRITERIA",
+						"Type" : "string",
+						"Optional" : true,
+                        "PermittedValues": [
+                              "AIC", 
+							  "AICC", 
+							  "BIC"
+                        ],
+						"DefaultValue": "AIC",
+                        "Description": [
+                                "Information criterion to be used in model selection. Options are AIC, AICc, BIC."
+                        ]
+                },
+				{
+                        "Name" : "STEPWISE",
+						"Type" : "integer",
+						"Optional" : true,
+                        "PermittedValues": [
+                               0,
+							   1
+                        ],
+						"DefaultValue": 0,
+                        "Description": [
+                                "If true, then the function does stepwise selection. If false, the the function selects all models. A value of 1 means true. A value of 0 means false."
+                        ]
+                },
+				{
+                        "Name" : "NMODELS",
+						"Type" : "integer",
+						"Optional" : true,
+                        "LowerBound": 1,
+						"UpperBound": 94,
+                        "LowerBoundType": "INCLUSIVE",
+                        "UpperBoundType": "INCLUSIVE",
+						"DefaultValue": 94,
+                        "Description": [
+                                "Maximum number of models considered in the stepwise search."
+                        ]
+                },
+				{
+                        "Name" : "MAX_ITERATIONS",
+						"Type" : "integer",
+						"Optional" : true,
+                        "LowerBound": 1,						
+                        "LowerBoundType": "INCLUSIVE",
+						"DefaultValue": 100,
+                        "Description": [
+                                "The maximum number of iterations that can be employed to non-linear optimization procedure."
+                        ]
+                },
+				{
+                        "Name" : "COEFF_STATS",
+						"Type" : "integer",
+						"Optional" : true,
+                        "PermittedValues": [
+                               0,
+							   1
+                        ],
+						"DefaultValue": 0,
+                        "Description": [
+                                "Indicator to return coefficient statistical columns TSTAT_VALUE and TSTAT_PROB. A value of 1 means return the columns. A value of 0 means do not return the columns."
+                        ]
+                },
+				{
+                        "Name" : "FIT_METRICS",
+						"Type" : "integer",
+						"Optional" : true,
+                        "PermittedValues": [
+                               0,
+							   1
+                        ],
+						"DefaultValue": 0,
+                        "Description": [
+                                "Indicator to generate the secondary result set that contains the model metadata statistics. A value of 1 means generate the secondary result set. A value of 0 means do not generate the secondary result set."
+                        ]
+                },
+				{
+                        "Name" : "RESIDUALS",
+						"Type" : "integer",
+						"Optional" : true,
+                        "PermittedValues": [
+                               0,
+							   1
+                        ],
+						"DefaultValue": 0,
+                        "Description": [
+                                "Indicator to generate the tertiary result set that contains the model residuals. A value of 1 means generate the tertiary result set. A value of 0 means do not generate the tertiary result set."
+                        ]
+                },
+				{
+                        "Name" : "ARMA_ROOTS",
+						"Type" : "integer",
+						"Optional" : true,
+                        "PermittedValues": [
+                               0,
+							   1
+                        ],
+						"DefaultValue": 0,
+                        "Description": [
+                                "Indicator to generate the senary result set that contains the inverse AR and MA roots of result best model that TD_AUTOARIMA selected (the model in the primary output layer). There should be no inverse roots showing outside of the unit circle. A value of 1 means generate result set. A value of 0 means do not generate a result set."
+                        ]
+                },
+				{
+                        "Name" : "TEST_NONSEASONAL",
+						"Type" : "string",
+						"Optional" : true,
+                        "PermittedValues": [
+                               "ADF"
+                        ],
+						"DefaultValue": "ADF",
+                        "Description": [
+                                "Nonseasonal unit root test used to choose differencing number d.",
+								"TD_AUTOARIMA function only uses ADF test for nonseasonal unit root test."
+                        ]
+                },
+				{
+                        "Name" : "TEST_SEASONAL",
+						"Type" : "string",
+						"Optional" : true,
+                        "PermittedValues": [
+                               "OCSB"
+                        ],
+						"DefaultValue": "OCSB",
+                        "Description": [
+                                "Seasonal unit root test used to choose differencing number D.",
+								"TD_AUTOARIMA function only uses OCSB test for seasonal unit root test."
+                        ]
+                }
+        ],
+        "InputFmt": false,
+        "OutputFmt": [
+                {
+                        "Name": "INDEX_STYLE",
+                        "Type": "string",
+                        "Optional": true,
+                        "DefaultValue": "NUMERICAL_SEQUENCE",
+                        "PermittedValues": [
+                                "NUMERICAL_SEQUENCE",
+                                "FLOW_THROUGH"
+                        ],
+                        "Description": [
+                                "[Optional] User can choose between an index style of NUMERICAL_SEQUENCE or FLOW_THROUGH."								
+                        ],
+                        "LangName": "output_fmt_index_style"
+                }
+        ]
+}

teradataml/data/jsons/uaf/17.20/TD_BINARYMATRIXOP.json

@@ -0,0 +1,107 @@
+{
+        "FuncName": "TD_BINARYMATRIXOP",
+        "FuncDescriptionShort": "Perform a point-wise mathematical operation against two matrices having the same number of elements",
+        "FuncDescriptionLong": [
+                "There is a need in data science to be able to create a new data matrix that is the result of performing a point-wise mathematical operation against two matrices having the same number of elements.   The principal mathematical operations of interest being that of subtraction, addition, multiplication, and division.  It is called a point-wise operation because it performs the mathematical operation on one sample point at a time.  For example, if the mathematical operation being performed is subtraction, then for two matrices each with NxM sample points,  a new matrix is generated in which:  ResultMatrix1_1 =  MinuendMatrix1_1 - SubtrahendMatrix1_1 ; ResultMatrix2_1 =  MinuendMatrix2_1 - SubtrahendMatrix2_1 ; ResultMatrix3_1 =  MinuendMatrix3_1 - SubtrahendMatrix3_1 ;  ..., up to ResultMatrixN_M  = MinuendMatrixN_M  - SubtrahendMatrixN_M.   The result matrix/2D array produced by the TD_BINARYMATRIXOP is the same size, 'NxM', as its inputs and inherits the identifiers, MATRIX_ID(), in this case, from the minuend matrix.",
+                "Thus the TD_BINARYMATRIXOP takes two equally sized logical-runtime matrices as input - the first matrix referenced in the function call is referred to as being the 'primary matrix', whereas the second matrix referenced is the secondary matrix.  The result matrix always inherits the identifiers, MATRIX_ID() from the primary matrix.  Mathematically, the significance of being a primary matrix or secondary matrix is as follows:",
+                "For MATHOP(SUB), subtraction, the secondary matrix is subtracted from the primary matrix.",
+                "For MATHOP(ADD), addition, the secondary matrix is added to the primary matrix.",
+                "For MATHOP(MUL), multiply, the primary matrix is multiplied by the secondary matrix.",
+                "For MATHOP(DIV), divide, the primary matrix is divided by the secondary matrix.",
+                "The TD_BINARYMATRIXOP function can be configured to operate in one of three input modes - ONE2ONE, MANY2ONE, and MATCH. These modes determine the number of primary matrices and number of secondary matrices involved in the operation, as well as determining how the primary and secondary matrix will be matched together. See the INPUT_FMT section for more information.",
+                "Some Use Case Examples for these Operators include:",
+                "A use for the subtraction - MATHOP(SUB) - and addition - MATHOP(ADD) - operations is to respectively remove trends from the original matrix and to later restore the trends while engaged in forecasting with the model.  Some background for this is: In building statistical ARMA-style forecasting models, a necessary precondition is that the time matrix being modelled be stationary: stationary with respect to mean; stationary with respect to covariance; and, stationary with respect to variance.    Transforming a non-stationary matrix into a stationary matrix typically requires the completion of two main tasks: the first task requiring the data scientist to model the trend or periodicity inherent in the data, then use that model to generate a representative matrix; the second task being to execute an operator that forms a new matrix, formed by subtracting the just generated trend or periodicity matrix from the original matrix.  The TD_BINARYMATRIXOP function supplies just such an operator, it forms one or more new matrices by enabling the data scientist to perform a point-wise subtraction between the original matrix, aka, the 'minuend-matrix', and the matrix to be subtracted, aka, the 'subtrahend', so minuend minus subtrahend.",
+                "A use case for the multiplication - MATHOP(SUB) - comes with respect to applying a filter function against a time matrix.   To apply a filter function - low pass, band pass, high pass - to a time matrix in the time domain requires a complex operation known as conjugation.  In contrast, if one first were to take the Fourier Transform of the time matrix and the Fourier Transform of the filter matrix; transitioning both data matrices from the time domain over to the frequency domain, then while in the frequency domain, the filter could be applied simply by running the two Fourier result matrices through a point-wise multiplier to produce a filtered result in the frequency domain.  In other words, conjugation in the time domain (a complex process) is equivalent to point-wise multiplication in the frequency domain.  Once the multiplication operation is complete, the data scientist could then take the inverse transform of the multiplied result set, obtaining the filtered version of the time matrix.",
+                "Use case for TD_BINARYMATRIXOP - Removal of trends present within a time matrix.   One task frequently performed by the Forecast Model developer is the removal of a trend from a time matrix, prior to modelling that matrix.   There are several different approaches available to the data scientist for removing trends.   As inferred in the below diagram, most of those involve executing a function against the original matrix that produces a new matrix containing the “trend”.   This initial processing step is followed by another processing step in which the original matrix and “trend” matrix is run through a point-wise subtractor, which subtracts - for each sample time point - the trend matrix from the original, producing a third 'detrended' matrix.  In other words, given an original matrix and a 'trend' matrix with five sample points at times: T0, T1, T2, T2, T4; the subtractor performs, original(T0) - trend(T0); original(T1) - trend(T1); …; original(T4) - trend(T4); producing a detrended matrix with five sample points.   The net effect of this operation, see diagram below, is to create a new matrix that is “rotated down”, once the trend has been removed.   Forecast Modelling can then proceed on the 'detrended' matrix."
+        ],
+        "FunctionVersion": "...",
+        "FunctionCategory": "Data Preparation",
+        "JSONVersion": "1",
+        "FuncRName": "td_Binary_Matrix_Op",
+        "MaxInputFiles": 2,
+        "Input": [
+                {
+                        "Type": "MATRIX",
+                        "Description": [
+                                "This function takes two equally sized data matrices as inputs, performs the specified mathematical operation - add, subtract, multiply, divide - on a sample point by sample point basis, producing a result matrix, which is the result of performing the mathematical operation."
+                        ],
+                        "LangName": "data or object or newdata or ... --> Langauges team can work with UAF team to come up with this field",
+                        "Optional": false
+                },
+                {
+                        "Type": "MATRIX",
+                        "Description": [
+                                "secondary matrix of equal size to be operated on"
+                        ],
+                        "LangName": "data or object or newdata or ... --> Langauges team can work with UAF team to come up with this field",
+                        "Optional": false
+                }
+        ],
+        "Output": [
+                {
+                        "Type": "ART",
+                        "PrimaryLayer": true,
+                        "LayerName": "ARTPRIMARY",
+                        "ResultTableColumnTypes": [
+                                "<varies>",
+                                "<varies>",
+                                "float"
+                        ],
+                        "Description": [
+                                "The TD_BINARYMATRIXOP function returns only a primary result set.   There are no secondary or tertiary result sets produced by this function.  The primary result set can be retrieved by issuing a SELECT against the analytical table containing the results.",
+                                "<varies> data type and values dependent upon the OUTPUT_FMT(INDEX_STYLE()) declaration."
+                        ],
+                        "LangName": "data or object or newdata or ... --> Langauges team can work with UAF team to come up with this field"
+                }
+        ],
+        "IsPlottable": true,
+        "Params": [
+                {
+                        "Name": "MATHOP",
+                        "Type": "string",
+                        "Optional": false,
+                        "PermittedValues": [
+                                "SUB",
+                                "ADD",
+                                "MUL",
+                                "DIV"
+                        ],
+                        "Description": [
+                                "Dictates mathematical operation to be performed between the passed in primary and secondary matrices.  It is an operation, where:  result-matrix = primary-matrix MATHOP secondary-matrix.   Specifics of the operation depend on the passed in PAYLOAD(CONTENT(); meaning that the operation will be a real operation, a multivariate real operation, a complex number operation, or an amplitude phase operation. "
+                        ],
+                        "LangName": "math_op"
+                }
+        ],
+        "InputFmt": [
+                {
+                        "Name": "INPUT_MODE",
+                        "Type": "string",
+                        "Optional": false,
+                        "PermittedValues": [
+                                "ONE2ONE",
+                                "MANY2ONE",
+                                "MATCH"
+                        ],
+                        "Description": [
+                                "Specifies the input mode supported by the function."
+                        ],
+                        "LangName": "input_fmt_input_mode"
+                }
+        ],
+        "OutputFmt": [
+                {
+                        "Name": "INDEX_STYLE",
+                        "Type": "string",
+                        "Optional": true,
+                        "DefaultValue": "NUMERICAL_SEQUENCE",
+                        "PermittedValues": [
+                                "NUMERICAL_SEQUENCE",
+                                "FLOW_THROUGH"
+                        ],
+                        "Description": [
+                                "Specifies the INDEX_STYLE of the output format."
+                        ],
+                        "LangName": "output_fmt_index_style"
+                }
+        ]
+}