Baseline Classifier Benchmark

TL;DR: Compare Logistic Regression vs Random Forest on a small dataset (Iris). Includes accuracy, ROC-AUC, and confusion matrix.

What’s inside

• Load the Iris dataset from scikit-learn
• Train two classifiers: Logistic Regression and Random Forest
• Evaluate using Accuracy, ROC-AUC, and Confusion Matrix
• PM-style commentary on tradeoffs

How to run

Open the notebook in Google Colab:

Results

• Logistic Regression accuracy: ~95%
• Random Forest accuracy: ~97%
• Random Forest performs slightly better, but Logistic Regression is simpler and faster to train.
• Both models separate classes well, but Random Forest handles nonlinear boundaries better.

Screenshot

Next ideas

• Try more classifiers (SVM, KNN, Gradient Boosting)
• Add cross-validation instead of a single train/test split
• Plot feature importance from Random Forest