fit_dt_regressor#

er_evaluation.fit_dt_regressor(X, y, numerical_features=None, categorical_features=None, sample_weights=None, random_state=0, criterion='squared_error', **kwargs)[source]#

Fits a decision tree regressor model with optional preprocessing for numerical and categorical features.

Parameters:

  • X (numpy array or pandas DataFrame) – The input features.

  • y (numpy array or pandas Series) – The target values.

  • numerical_features (list of int or str, optional) – The column indices or column names of numerical features. Default is None.

  • categorical_features (list of int or str, optional) – The column indices or column names of categorical features. Default is None.

  • sample_weights (numpy array, optional) – Individual weights for each sample. Default is None.

  • random_state (int) – Random state for the decision tree regressor.

  • criterion (str) – The function to measure the quality of a split. Supported criteria are “squared_error”, “friedman_mse”, “absolute_error”, and “poisson”. Default is “squared_error”.

  • **kwargs – Additional keyword arguments passed to the DecisionTreeRegressor constructor.

Returns:

A fitted decision tree regressor model with preprocessing steps.

Return type:

sklearn.pipeline.Pipeline

Examples

>>> import numpy as np
>>> import pandas as pd
>>> X = pd.DataFrame({"A": [1, 2, 3], "B": [4, 5, 6], "C": ["a", "b", "a"]})
>>> y = np.array([2, 4, 6])
>>> numerical_features = ["A", "B"]
>>> categorical_features = ["C"]
>>> model = fit_dt_regressor(X, y, numerical_features, categorical_features)
>>> isinstance(model, Pipeline)
True