Confusion Matrix Calculator
Overview
The Confusion Matrix Calculator computes comprehensive classification performance metrics from your model predictions. Simply input your predicted and actual labels (or upload a CSV), and the tool generates a color-coded confusion matrix along with all standard metrics including accuracy, precision, recall, F1 score, and per-class breakdowns. It supports both binary and multi-class classification problems.
Tips
- Use the CSV upload option for large datasets rather than pasting thousands of labels manually
- For imbalanced datasets, focus on precision and recall for minority classes rather than overall accuracy
- The per-class metrics table helps identify which classes your model struggles with most
- Compare macro-averaged vs weighted-averaged metrics to understand performance across all classes equally vs proportionally
- For binary classification, check the MCC (Matthews Correlation Coefficient) for a single balanced metric
- Export results to CSV for tracking model performance over time or comparing different models
- Use the direct matrix input if you already have aggregated counts from another tool or system