PCA_Analysis.ipynb

https://github.com/rhyslwells/ML_Tools/blob/main/Explorations/Preprocess/PCA/PCA_Analysis.ipynb

This script performs Principal Component Analysis (PCA) on the Iris dataset to reduce its dimensionality while preserving key variance.

See also PCA

Summary:

1. Load and Preprocess Data

   • Loads the Iris dataset and extracts features and target labels.
   • Scales the data to standardize feature ranges.

2. Apply PCA (3 Components)

   • Fits PCA to the scaled data and transforms it into three principal components.
   • Stores the transformed data in a DataFrame with species labels.

3. Analyze PCA Loadings & Variance

   • Computes and stores PCA loadings (weights of original features in principal components).
   • Computes explained variance and cumulative variance to assess PCA effectiveness.

4. Visualizations

   • Explained variance: Bar plot of individual and cumulative variance contributions.
   • PCA Scores: 3D scatter plots of transformed data, colored by species.
   • PCA Loadings: 3D scatter plot showing feature contributions to principal components.
   • Heatmap: Displays PCA component weights for feature importance analysis.

5. Additional Full PCA Analysis

   • Computes and prints explained variance for all components.
   • Uses Seaborn to generate a heatmap of PCA component contributions.