chore: clean up repository

Author: vxa8502

.gitignore

@@ -64,10 +64,5 @@ personal/
 # Test scripts in root
 test_*.sh
 
-# Generated results (metrics in README)
-results/
-
 # Dev-only scripts
-scripts/reset.sh
 scripts/cleanup.sh
-scripts/generate_report.sh

results/FINAL_COMPREHENSIVE_REPORT.md

@@ -0,0 +1,145 @@
+# Sentiment Analysis: Final Comprehensive Evaluation Report
+
+**Project**: Cross-Domain Sentiment Classification
+**Date**: January 30, 2026
+**Author**: Victoria Alabi
+**Generated**: Auto-generated from model metadata (do not edit manually)
+
+---
+
+## Executive Summary
+
+This report summarizes the training and evaluation of sentiment analysis models across multiple algorithms and domains.
+
+### Production Model
+
+**Algorithm**: SVM
+**Training Dataset**: amazon_polarity (40000 samples)
+**Model Size**: 15.9 MB
+
+| Metric | Value |
+|--------|-------|
+| Test Accuracy | 89.6% |
+| Test F1 | 0.896 |
+| Test Precision | 0.896 |
+| Test Recall | 0.896 |
+| ROC-AUC | 0.957 |
+| Cross-Domain Avg | 88.0% |
+
+### Production Model Confusion Matrix
+
+| | Predicted Negative | Predicted Positive |
+|---|---|---|
+| **Actual Negative** | 4478 (TN) | 525 (FP) |
+| **Actual Positive** | 519 (FN) | 4478 (TP) |
+
+### Best Generalizing Model
+
+**Model**: svm-amazon_polarity
+**Cross-Domain Average Accuracy**: 88.0%
+
+---
+
+## Part 1: Model Comparison (All 12 Experiments)
+
+| Algorithm | Dataset | Accuracy | F1 | Precision | Recall | Training Time |
+|-----------|---------|----------|-----|-----------|--------|---------------|
+| SVM | amazon_polarity | 89.2% | 0.892 | 0.893 | 0.892 | 97m 22s |
+| SVM | imdb_50k | 89.1% | 0.891 | 0.891 | 0.891 | 245m 56s |
+| SVM | yelp | 94.0% | 0.940 | 0.940 | 0.940 | 22m 32s |
+| LOGISTIC_REGRESSION | amazon_polarity | 84.1% | 0.841 | 0.841 | 0.841 | 53m 28s |
+| LOGISTIC_REGRESSION | imdb_50k | 85.5% | 0.855 | 0.855 | 0.855 | 87m 57s |
+| LOGISTIC_REGRESSION | yelp | 92.6% | 0.926 | 0.926 | 0.926 | 24m 48s |
+| RANDOM_FOREST | amazon_polarity | 86.6% | 0.866 | 0.866 | 0.866 | 101m 40s |
+| RANDOM_FOREST | imdb_50k | 86.6% | 0.866 | 0.866 | 0.866 | 223m 51s |
+| RANDOM_FOREST | yelp | 92.0% | 0.920 | 0.921 | 0.920 | 37m 47s |
+| NAIVE_BAYES | amazon_polarity | 79.2% | 0.792 | 0.796 | 0.792 | 59m 38s |
+| NAIVE_BAYES | imdb_50k | 82.9% | 0.829 | 0.830 | 0.829 | 115m 9s |
+| NAIVE_BAYES | yelp | 81.3% | 0.813 | 0.816 | 0.813 | 10m 39s |
+
+---
+
+## Part 2: Cross-Domain Evaluation
+
+Each model was evaluated on all three test domains. Asterisk (*) indicates in-domain evaluation.
+
+#### SVM
+
+| Train Domain | IMDB Test | Amazon Test | Yelp Test | Cross-Domain Avg |
+|--------------|-----------|-------------|-----------|------------------|
+| imdb 50k | 89.1% * | 81.9% | 84.9% | 83.4% |
+| amazon polarity | 85.2% | 89.2% * | 90.9% | 88.0% |
+| yelp | 78.5% | 82.0% | 94.0% * | 80.3% |
+
+#### LOGISTIC REGRESSION
+
+| Train Domain | IMDB Test | Amazon Test | Yelp Test | Cross-Domain Avg |
+|--------------|-----------|-------------|-----------|------------------|
+| imdb 50k | 85.5% * | 76.8% | 79.3% | 78.1% |
+| amazon polarity | 80.4% | 84.1% * | 84.7% | 82.6% |
+| yelp | 75.8% | 77.8% | 92.6% * | 76.8% |
+
+#### RANDOM FOREST
+
+| Train Domain | IMDB Test | Amazon Test | Yelp Test | Cross-Domain Avg |
+|--------------|-----------|-------------|-----------|------------------|
+| imdb 50k | 86.6% * | 80.2% | 83.1% | 81.6% |
+| amazon polarity | 79.6% | 86.6% * | 89.6% | 84.6% |
+| yelp | 71.5% | 81.3% | 92.0% * | 76.4% |
+
+#### NAIVE BAYES
+
+| Train Domain | IMDB Test | Amazon Test | Yelp Test | Cross-Domain Avg |
+|--------------|-----------|-------------|-----------|------------------|
+| imdb 50k | 82.9% * | 73.5% | 79.1% | 76.3% |
+| amazon polarity | 69.4% | 79.2% * | 81.3% | 75.4% |
+| yelp | 59.8% | 70.9% | 81.3% * | 65.3% |
+
+**Legend**: * = in-domain evaluation
+
+---
+
+## Part 3: Reproducibility
+
+All results can be reproduced via:
+
+```bash
+# Prepare immutable data splits (run once)
+./scripts/prepare_data.sh
+
+# Train all models
+./scripts/train_all_models.sh
+
+# Cross-domain evaluation
+./scripts/evaluate_cross_domain.sh
+
+# Regenerate this report
+./scripts/generate_report.sh
+```
+
+### Model Artifacts
+
+| File | Algorithm | Dataset | Size |
+|------|-----------|---------|------|
+| amazon_polarity_logistic_regression_model.ser | LOGISTIC_REGRESSION | amazon_polarity | 14.4 MB |
+| imdb_50k_logistic_regression_model.ser | LOGISTIC_REGRESSION | imdb_50k | 30.1 MB |
+| yelp_logistic_regression_model.ser | LOGISTIC_REGRESSION | yelp | 9.7 MB |
+| amazon_polarity_naive_bayes_model.ser | NAIVE_BAYES | amazon_polarity | 14.5 MB |
+| imdb_50k_naive_bayes_model.ser | NAIVE_BAYES | imdb_50k | 30.3 MB |
+| yelp_naive_bayes_model.ser | NAIVE_BAYES | yelp | 9.9 MB |
+| sentiment_model.ser | SVM | amazon_polarity | 15.9 MB |
+| amazon_polarity_random_forest_model.ser | RANDOM_FOREST | amazon_polarity | 194.6 MB |
+| imdb_50k_random_forest_model.ser | RANDOM_FOREST | imdb_50k | 193.7 MB |
+| yelp_random_forest_model.ser | RANDOM_FOREST | yelp | 91.9 MB |
+| amazon_polarity_svm_model.ser | SVM | amazon_polarity | 15.9 MB |
+| imdb_50k_svm_model.ser | SVM | imdb_50k | 31.6 MB |
+| yelp_svm_model.ser | SVM | yelp | 10.3 MB |
+
+---
+
+## Metadata
+
+- **Report Generated**: 2026-01-30T20:40:18Z
+- **Git Commit**: 6d7f5e5
+- **Java Version**: 24.0.1
+