Fit Fast, Explain Fast
Project description
# FastExplain > Fit Fast, Explain Fast
## Installing ` pip install fast-explain ` ## Clean Data, Fit ML Models and Explore Results all in one line. FastExplain provides an out-of-the-box tool for analysts to quickly model and explore data, with flexibility to fine-tune if needed. - Automated cleaning and fitting of machine learning models with hyperparameter search - Aesthetic display of explanatory methods ready for reporting - Connected interface for all data, models and related explanatory methods
## Quickstart
### Automated Cleaning and Fitting ` python from FastExplain import * df = load_titanic_data() classification = model_data(df, dep_var="Survived", model="rf", hypertune=True) ` ### Aesthetic Display ` python plot_one_way_analysis(classification.data.df, "Age", "Survived", filter = "Sex == 1") ` <img alt=”One Way” src=”images/one_way.png”>
` python plot_ale(classification.m, classification.data.xs, "Age", filter = "Sex == 1", dep_name = "Survived") ` <img alt=”ALE” src=”images/ALE.png”>
` python classification_1 = model_data(df, dep_var="Survived", model="rf", hypertune=True, cont_names=['Age'], cat_names = []) models = [classification.m, classification_1.m] data = [classification.data.xs, classification_1.data.xs] plot_ale(models, data, 'Age', dep_name = "Survived") ` <img alt=”multi_ALE” src=”images/multi_ALE.png”>
### Connected Interface ` python classification.plot_one_way_analysis("Age", filter = "Sex == 1") classification.plot_ale("Age", filter = "Sex == 1") `
` python classification.shap_dependence_plot("Age", filter = "Sex == 1") ` <img alt=”SHAP” src=”images/shap.png”>
` python classification.error # {'auc': {'model': {'train': 0.9934332941166654, # 'val': 0.8421607378129118, # 'overall': 0.9665739941840028}}, # 'cross_entropy': {'model': {'train': 0.19279692001978943, # 'val': 0.4600233891109683, # 'overall': 0.24648214781700722}}} `
## Models Supported - Random Forest - XGBoost - Explainable Boosting Machine - ANY Model Class with fit and predict attributes
## Exploratory Methods Supported: - One-way Analysis - Two-way Analysis - Feature Importance Plots - ALE Plots - Explainable Boosting Methods - SHAP Values - Partial Dependence Plots - Sensitivity Analysis
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distributions
Built Distribution
Hashes for fast_explain-0.0.46-py2.py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | e4ed765a013d4e5c3876b534ad5ea5d8300160fcd99ea186ef156ad463f43178 |
|
MD5 | 40fcb3ca67521a93aca57093829b850f |
|
BLAKE2b-256 | 2db6a16e805f9a7d3a4a0daea9be42968a1c59b5d19af9e2f59f6d6ab8c1960e |