ourtils.modeling

Contents

ourtils.modeling#

All things modeling

class ourtils.modeling.RegResultCollection(reg_results: list[RegressionResult])#

Bases: object

Represents a collection of regression models.

batch_plot_regs(x_var: str, n=10)#

classmethod build_from_mapper(input_data: DataFrame, model_mapper: list[tuple])#

Creates a RegResultCollection from a dictionary

model_mapper: A dictionary of: ```python import pandas as pd {

(key, formula): model description

}#

input_data: The input dataframe to fit the model on

coef_dataframe(*args, **kwargs) → DataFrame#: Returns model summaries as a dataframe

property model_summary#

plot_coefs(show_nums=True, scale_by=None, *args, **kwargs)#: Plots the coefficients of each model as a bar chart with error bars.

property rowlevel_data: DataFrame#

class ourtils.modeling.RegressionResult(key: str, formula: str, description: str, data: DataFrame)#

Bases: object

Represents a single model

property diagnostic_df: DataFrame#

property fitted_model: RegressionResultsWrapper#

get_n_most_influential_points(n=5)#

get_param_df(standardize=True) → DataFrame#

idx_most_influential(n)#

property info_df: DataFrame#

property model_descr#

plot_coefs(*args, **kwargs)#

plot_leverage_vs_resids()#

plot_n_influential(x_val: str, n: int, title=None, ax=None) → None#: Plots the n most influential points.

plot_partial_regression()#

plot_qq()#

plot_scatter(x_val: str, y_val=None, highlight_influence=False, **kwargs)#

property regression_influences#

property standardized_fitted_model#: Standardizes data before fitting.

property summary#

tag_dataframe(_df) → DataFrame#

property y_variable: str#

class ourtils.modeling.Sklearner(X, y, preprocessing_pipeline, models: list)#

Bases: object

create_cv_results(*args, **kwargs) → DataFrame#

ourtils.modeling.standardize(numbers: Series) → Series#: Safely standardizes a series.