🎯 Model Accuracy Predictor
Enter the four cells of your confusion matrix to see accuracy, precision, recall, F1, and specificity — the metrics that reveal how a classifier truly performs.
🧮 Evaluate Your Classifier
What is a Model Accuracy Predictor?
A single accuracy number rarely tells the whole story of a machine-learning model. This tool takes the four counts from a confusion matrix and works out the family of metrics data scientists rely on — precision, recall, F1, and specificity — so you can see the trade-offs behind the headline figure.
It's built for learning: change the true and false positive and negative counts and watch each metric respond, and you'll quickly grasp why accuracy alone can hide a model that's failing on the class you care about most.
❓ Frequently Asked Questions
What is a confusion matrix?
It's a small table that summarises a binary classifier's results across four cells: true positives (correctly predicted positive), true negatives (correctly predicted negative), false positives (predicted positive but actually negative — a false alarm), and false negatives (predicted negative but actually positive — a miss). Every metric on this page is computed directly from those four counts, so it's the natural starting point for evaluating a model.
What's the difference between precision and recall?
Precision asks: of everything the model flagged as positive, how many really were? It's TP ÷ (TP + FP). Recall asks: of everything that truly was positive, how many did the model catch? It's TP ÷ (TP + FN). A spam filter with high precision rarely flags real mail as spam; one with high recall rarely lets spam through. You usually trade one against the other, which is exactly why the F1 score exists.
What is the F1 score and why not just use accuracy?
Accuracy is the share of all predictions that were correct, but it can be badly misleading on imbalanced data — if 99% of examples are negative, a model that always predicts negative scores 99% accuracy while catching zero positives. The F1 score is the harmonic mean of precision and recall, so it stays low unless both are reasonably high. This tool computes F1 from the rounded precision and recall so the displayed numbers reconcile.
Can this predict my model's accuracy before training?
No — it measures accuracy from results you already have. You run your trained model on a labeled test set, count the four confusion-matrix cells, and enter them here to see the full metric picture. It's an evaluation and learning tool: the fastest way to understand what precision, recall, F1, and specificity mean and how they move as the counts change.