Outlier_treatment-and-Data_transformation-using-various-methods

This repository focuses on statistical techniques for handling missing values, detecting and treating outliers, and transforming data distributions to improve data quality for machine learning models.

Concepts Covered

• Identifying missing values
• Filling missing data using:
  • Mean
  • Median
  • Mode
• Outlier detection using IQR method
• Removing outliers using boolean masking (~)
• Handling outliers across multiple numerical columns
• Capping / Winsorization (1st–99th percentile)
• Data transformation techniques:
  • Log transformation
  • Square root / Power transformation (Yeo-Johnson)
  • Box-Cox transformation
• Scaling
• Normalizing

Libraries Used

• pandas, numpy
• scikit-learn
• scipy

🎯 Why This Matters

Outlier treatment and data transformation help stabilize variance, reduce skewness, and improve the performance of machine learning algorithms.

---

⭐ Star the repository if you find it useful!