constants.py

TD_ANALYTIC_FUNCS = {
    "ANOVA": "Performs one-way Analysis of Variance (ANOVA) on a dataset with two or more groups.",
    "Attribution": "Assigns weights to web pages before conversion events to measure their influence on outcomes like product purchases.",
    "Antiselect": "Returns all columns from the input table except those specified in the Exclude argument.",
    "Apriori": "Finds association patterns and calculates statistical metrics to understand the influence of item sets on each other.",
    "BincodeFit": "Computes bin boundaries for numeric columns to be applied by BincodeTransform().",
    "BincodeTransform": "Converts continuous numeric data to categorical data using bin boundaries from BincodeFit() output.",
    "CFilter": "Calculates statistical measures of how likely each pair of items is to be purchased together.",
    "CategoricalSummary": "Displays distinct values and their counts for each specified input column.",
    "ChiSq": "Performs Pearson's chi-squared test for independence between two categorical variables.",
    "ClassificationEvaluator": "Evaluates a classification model by computing confusion matrix metrics such as accuracy, precision, recall, and F1.",
    "ColumnSummary": "Provides a quick overview of column datatypes and a summary of NULL and non-NULL counts for a given table.",
    "ColumnTransformer": "Applies multiple transformations to input data columns in a single operation using Fit analytic function outputs.",
    "ConvertTo": "Converts specified input columns to specified data types.",
    "DecisionForest": "Trains an ensemble decision forest model for classification and regression predictive modeling.",
    "FTest": "Performs an F-test where the test statistic follows an F-distribution under the null hypothesis.",
    "FillRowId": "Adds a column of unique row identifiers to the input table.",
    "Fit": "Determines whether specified numeric transformations can be applied to target columns and outputs parameters for Transform().",
    "GetFutileColumns": "Returns names of columns that are futile — all values unique, all values identical, or distinct ratio exceeds a threshold.",
    "GetRowsWithMissingValues": "Returns rows that contain NULL values in any of the specified input columns.",
    "GetRowsWithoutMissingValues": "Returns rows that have non-NULL values in all of the specified input columns.",
    "GLM": "Trains a Generalized Linear Model (GLM) for regression and classification on data following an exponential family distribution.",
    "GLMPerSegment": "Trains a separate GLM model for each segment of the input data.",
    "Histogram": "Calculates frequency distribution of a dataset using Sturges, Scott, variable-width, or equal-width binning methods.",
    "KMeans": "Groups observations into k clusters where each point belongs to the cluster with the nearest centroid.",
    "KMeansPredict": "Assigns input data points to cluster centroids produced by KMeans().",
    "KNN": "Classifies data points based on proximity to training data points with known categories.",
    "MovingAverage": "Computes moving average values in a series using a specified moving average type.",
    "NERExtractor": "Performs Named Entity Recognition (NER) on input text using dictionary words or regular expression patterns.",
    "NGramSplitter": "Tokenizes an input stream of text and outputs n-grams based on specified delimiter and reset parameters.",
    "NaiveBayesTextClassifierPredict": "Predicts text categories using a model generated by NaiveBayesTextClassifierTrainer().",
    "NaiveBayesTextClassifierTrainer": "Calculates conditional probabilities and prior probabilities for token-category pairs for text classification.",
    "NonLinearCombineFit": "Computes parameters for a non-linear combination of existing features for use by NonLinearCombineTransform().",
    "NonLinearCombineTransform": "Generates a new feature by applying a non-linear combination formula using NonLinearCombineFit() output.",
    "NumApply": "Applies a predefined numeric operation to specified input columns.",
    "NPath": "Scans a set of rows looking for user-specified sequential patterns and returns rows that match.",
    "OneClassSVM": "Trains a linear one-class SVM model to identify outliers or novelty in a dataset.",
    "OneClassSVMPredict": "Predicts whether input data points are outliers using a model generated by OneClassSVM().",
    "OneHotEncodingFit": "Identifies categorical values to be encoded and outputs parameters for OneHotEncodingTransform().",
    "OneHotEncodingTransform": "Encodes categorical columns as one-hot numeric vectors using OneHotEncodingFit() output.",
    "OrdinalEncodingFit": "Identifies distinct categorical values and generates ordinal mappings for use with OrdinalEncodingTransform().",
    "OrdinalEncodingTransform": "Maps categorical values to ordinal integers using OrdinalEncodingFit() output.",
    "OutlierFilterFit": "Calculates percentile bounds and median for target columns for use by OutlierFilterTransform().",
    "OutlierFilterTransform": "Filters rows containing outlier values using bounds from OutlierFilterFit() output.",
    "Pack": "Packs data from multiple input columns into a single column.",
    "PolynomialFeaturesFit": "Computes polynomial combination parameters for existing features for use by PolynomialFeaturesTransform().",
    "PolynomialFeaturesTransform": "Generates polynomial feature combinations from existing features using PolynomialFeaturesFit() output.",
    "Pivoting": "Pivots data from sparse format to dense format (rows to columns).",
    "QQNorm": "Determines whether values in input columns follow a normal distribution.",
    "ROC": "Computes true positive rate, false positive rate, AUC, and Gini coefficient for a binary classification model across thresholds.",
    "RandomProjectionFit": "Generates a random projection matrix based on specified arguments for use by RandomProjectionTransform().",
    "RandomProjectionMinComponents": "Calculates the minimum number of components required for random projection given an epsilon distortion value.",
    "RandomProjectionTransform": "Reduces high-dimensional input data to a lower-dimensional space using RandomProjectionFit() output.",
    "RegressionEvaluator": "Computes metrics to evaluate regression model predictions including RMSE, MAE, and R-squared.",
    "RoundColumns": "Rounds values in specified input columns to a specified number of decimal places.",
    "RowNormalizeFit": "Computes row-wise normalization parameters for specified columns for use by RowNormalizeTransform().",
    "RowNormalizeTransform": "Normalizes input columns row-wise using RowNormalizeFit() output.",
    "SMOTE": "Generates synthetic minority class samples using SMOTE, ADASYN, Borderline-2, or SMOTE-NC algorithms.",
    "SVM": "Trains a linear Support Vector Machine (SVM) for classification and regression.",
    "SVMPredict": "Predicts target values or class labels on new data using a model generated by SVM().",
    "ScaleFit": "Computes scaling statistics for specified columns for use by ScaleTransform().",
    "ScaleTransform": "Scales specified columns using statistics from ScaleFit() output.",
    "Sessionize": "Maps each click event in a user session to a unique session identifier.",
    "SentimentExtractor": "Extracts the sentiment (positive, negative, or neutral) of each input document or sentence using a dictionary model.",
    "Shap": "Computes Shapley values to explain individual predictions (feature contributions) for a machine learning model.",
    "Silhouette": "Measures the consistency of cluster assignments by computing silhouette scores for each data point.",
    "SimpleImputeFit": "Computes imputation values (mean, median, or mode) for missing values in the input data.",
    "SimpleImputeTransform": "Substitutes missing values in the input data using imputation values from SimpleImputeFit() output.",
    "StrApply": "Applies a predefined string operation to specified input columns.",
    "StringSimilarity": "Calculates similarity between two strings using Jaro, Jaro-Winkler, N-Gram, or Levenshtein distance.",
    "TDDecisionForestPredict": "Predicts target values or class labels using a DecisionForest() model.",
    "TDGLMPredict": "Predicts target values or class labels for test data using a GLM() model.",
    "TDNaiveBayesPredict": "Predicts classification labels using a model generated by NaiveBayes().",
    "TFIDF": "Computes Term Frequency (TF), Inverse Document Frequency (IDF), and TF-IDF scores for each term in a document set.",
    "TargetEncodingFit": "Computes target encoding values (expected value per category) for categorical columns.",
    "TargetEncodingTransform": "Encodes categorical columns using encoding values from TargetEncodingFit() output.",
    "TextMorph": "Generates morphological variants (morphs) of words in the input dataset.",
    "TextParser": "Tokenizes text, removes punctuation, converts to lowercase, removes stopwords, and applies stemming or lemmatization.",
    "TrainTestSplit": "Splits input data into training and test sets to simulate model performance on new data.",
    "Transform": "Applies numeric transformations to input columns using parameters from Fit() output.",
    "UnivariateStatistics": "Displays descriptive statistics (mean, min, max, stddev, etc.) for each specified numeric input column.",
    "Unpack": "Unpacks data from a single packed column into multiple separate columns.",
    "Unpivoting": "Unpivots data from dense format to sparse format (columns to rows).",
    "VectorDistance": "Computes distances between target vectors and reference vectors.",
    "WhichMax": "Returns all rows that contain the maximum value in a specified input column.",
    "WhichMin": "Returns all rows that contain the minimum value in a specified input column.",
    "WordEmbeddings": "Produces embedding vectors for text and computes similarity between texts.",
    "XGBoost": "Trains an XGBoost (eXtreme Gradient Boosting) model for classification or regression.",
    "XGBoostPredict": "Predicts target values or class labels using a model generated by XGBoost().",
    "ZTest": "Tests the equality of two means under the assumption that the population variances are known.",
}