automl-tools 0.2.5

automl_tools

Automl_tools: automl binary classification

Automl_tools is a Python library that implements Gradient Boosting

Installation

The code is packaged for PyPI, so that the installation consists in running:

pip install automl-tools

Colab

Usage

Probabilistic binary example on the Boston housing dataset:

import pandas as pd
from automl_tools import automl_run

train = pd.read_csv("https://raw.githubusercontent.com/jonaqp/automl_tools/main/automl_tools/examples/train.csv?token=AAN2ZBDWF77QITK4ARSFIFDABUGAU")
test = pd.read_csv("https://raw.githubusercontent.com/jonaqp/automl_tools/main/automl_tools/examples/test.csv?token=AAN2ZBD6TMUC5XSGRTJNVPDABUGCO")

automl_run(train=train,
           test=test,
           id_col=None, 
           target_col="Survived",
           imp_num="knn",
           imp_cat="knn",
           processing="binding",
           mutual_information=False,
           correlation_drop=False,
           model_feature_selection=None,
           model_run="LR",
           augmentation=True,
           Stratified=True,
           cv=5)

Parameter

imp_num : "gaussian", "arbitrary", "median", "mean", "random", "knn"
imp_cat : "frequent", "constant", "rare", "knn"
processing:  "woe", "binding" 

Support Binary

model_feature_selection: 
    default: ["LR", "RF", "LGB"]
        LR  : LogisticRegression
        RF  : RandomForestClassifier
        SVM : SVC
        LS  : LASSO
        RD  : RIDGE
        NET : Elasticnet
        DT  : DecisionTreeClassifier
        ET  : ExtraTreesClassifier
        GB  : GradientBoostingClassifier
        AB  : AdaBoostClassifier
        XGB  : XGBClassifier
        LGB  : LGBMClassifier
        CTB  : CatBoostClassifier
        NGB  : NGBClassifier

model_run:
    default: "LR"
        LR  : LogisticRegression
        RF  : RandomForestClassifier
        SVM : SVC
        LS  : LASSO
        RD  : RIDGE
        NET : Elasticnet
        DT  : DecisionTreeClassifier
        ET  : ExtraTreesClassifier
        GB  : GradientBoostingClassifier
        AB  : AdaBoostClassifier
        XGB  : XGBClassifier
        LGB  : LGBMClassifier
        CTB  : CatBoostClassifier
        NGB  : NGBClassifier

License

Apache License 2.0.

New features v1.0

• multi_class
• regression
• integrations GCP deploy model CI/CD
• integrations AWS deploy model CI/CD

BugFix

• 0.1.5

  • fix imputer
  • fix space hyperparameter
  • update catboost test

• 0.1.4

  • add parameter cv
  • add confusion Matrix
  • add comments readme.txt

• 0.1.3

  • add parameter id_col
  • add comments readme.txt

Reference

• Jonathan Quiza github.
• Jonathan Quiza RumiMLSpark.
• Jonathan Quiza linkedin.