AutoFEAT

Framework for automated feature generation based on statistics and AI tools

Installation

pip install autofe-grass

Quick Start

import pandas as pd
from autofe import GroupAggregationFeatures, StatisticalFeatureGenerator

# Load your data
df = pd.read_csv('your_data.csv')
X_train, X_test, y_train, y_test = train_test_split(X, y)

# Generate group-based features
group_feats1 = GroupAggregationFeatures(
        numeric_cols=numeric_cols,
        group_cols=group_cols,
        aggs=['mean', 'std' ...],  
        add_deviations=True,
        add_rank=False
)
X_train_grouped = group_features.fit_transform(X_train)

# Generate statistical features
stat_gen = StatisticalFeatureGenerator(
        numeric_cols=numeric_cols,
        unary=['log', 'sqrt'],
        pairwise=['ratio', 'diff'],
        max_features=20,
        corr_th=0.95,
        min_var=1e-5
    )
X_train_stats = stat_features.fit_transform(X_train_grouped, y_train)

Key Features

• Group Aggregations - Mean, std, min, max, sum, count by categories
• Statistical Transforms - Log, sqrt, ratio, difference between features
• Sklearn-Compatible - Works with sklearn Pipeline

🔧 Requirements

• Python >= 3.8
• numpy >= 1.19.0
• pandas >= 1.2.0
• scikit-learn >= 0.24.0