autopreprocess-lite 0.1.1

Automatic Data Preprocessing Library

AutoPreprocess

Automatic Data Preprocessing Library for Machine Learning

✨ Features

• Automatic column type detection (numeric, categorical, datetime, useless)
• Smart missing value handling (based on missing percentage)
• Outlier detection & capping (IQR method)
• Intelligent encoding (One-hot, Frequency, Target encoding)
• Feature scaling (Standard, MinMax, Robust)
• Feature selection (Variance, Correlation, Importance, Mutual Info)
• Train/Test split (Random, Stratified, Time series)
• Save & load pipeline for production deployment
• Zero data leakage (fit only on training data)

🚀 Quick Start

from autopreprocess import AutoClean

# One line to preprocess everything
pipeline = AutoClean('data.csv', target='price')
X_train, X_test, y_train, y_test = pipeline.preprocess()

# For new predictions
X_new_clean = pipeline.predict_ready_data(X_new)

# Save for later
pipeline.save('my_pipeline.pkl')

# Load and use
loaded = AutoClean.load('my_pipeline.pkl')