← back to projects

/// project

Wine Classification

Wine classifier pipeline using Logistic Regression

  • Python
  • Machine Learning
  • Logistic Regression
  • PCA
  • Built an end-to-end ML pipeline on the 178-sample, 13-feature Wine dataset using train/test splitting and StandardScaler normalization to prepare data for PCA and classification.
  • Interpreted principal components via loading heatmaps and biplots, identifying Flavanoids and OD280 as top PC1 drivers and Color Intensity and Alcohol as top PC2 drivers.
  • Benchmarked Logistic Regression, SVM (RBF), and KNN (k=5) on PCA-reduced features, achieving 97.2% test accuracy and ˜97% 10-fold CV mean accuracy across all three models.