johnson transformation

**Johnson transformation** is the **flexible distribution-mapping approach that converts a wide range of non-normal data shapes into near-normal form** - it is useful when simpler power transforms cannot adequately normalize complex distributions. **What Is Johnson transformation?** - **Definition**: Parametric family of SL, SU, and SB transformations selected to match data shape and bounds. - **Strength**: Can model bounded, unbounded, and highly skewed distributions more flexibly than single-power methods. - **Use Case**: Applied when Box-Cox fit quality is insufficient for reliable capability estimation. - **Output**: Transformed values suitable for normal-based probability and capability calculations. **Why Johnson transformation Matters** - **Fit Flexibility**: Handles complex shapes that appear in mixed physical mechanisms and bounded responses. - **Tail Accuracy**: Better transformation fit improves defect-rate prediction near critical limits. - **Method Continuity**: Allows teams to keep standard SPC infrastructure while addressing non-normal data. - **Decision Quality**: Reduces risk of false acceptance from poorly fitted normal assumptions. - **Advanced Analytics**: Supports robust capability reporting in challenging manufacturing datasets. **How It Is Used in Practice** - **Family Selection**: Choose Johnson family variant through goodness-of-fit optimization. - **Parameter Estimation**: Fit transformation constants and validate with transformed normality diagnostics. - **Capability Reporting**: Compute transformed-domain indices and provide clear interpretation guidance. Johnson transformation is **the high-flexibility option for difficult non-normal capability problems** - when simpler methods fail, Johnson mapping often restores usable statistical inference.