Expand description
Adaptive preprocessing parameters that automatically tune based on data characteristics
This module provides automatic parameter selection and tuning for preprocessing transformers based on statistical analysis of the input data. It helps optimize preprocessing pipelines without manual parameter tuning.
§Features
- Data Distribution Analysis: Automatically detect data distribution characteristics
- Adaptive Thresholds: Dynamic threshold selection based on data properties
- Parameter Optimization: Automatic parameter tuning for various transformers
- Multi-Objective Optimization: Balance multiple criteria (robustness, efficiency, quality)
- Cross-Validation Based Tuning: Use CV to select optimal parameters
- Ensemble Parameter Selection: Combine multiple parameter selection strategies
Structs§
- Adaptive
Config - Configuration for adaptive parameter selection
- Adaptive
Parameter Selector - Adaptive parameter selector for preprocessing transformers
- Data
Characteristics - Data distribution characteristics detected from input data
- Imputation
Parameters - Adaptive imputation parameters
- Outlier
Detection Parameters - Adaptive outlier detection parameters
- Parameter
Evaluation - Parameter evaluation result
- Parameter
Recommendations - Adaptive parameter recommendations for different transformers
- Scaling
Parameters - Adaptive scaling parameters
- Transformation
Parameters - Adaptive transformation parameters
Enums§
- Adaptation
Strategy - Adaptive parameter selection strategies
- Distribution
Type - Types of distributions detected in features