Evaluates the performance of a classification model using metrics derived from a confusion matrix, including accuracy, precision, recall, and specificity.
Start with your values, or try a preset.
Enter True Positives (TP), False Positives (FP), True Negatives (TN), and False Negatives (FN) in the units shown.
Read Specificity, Recall (Sensitivity), Precision, Accuracy, and F1 Score in the results panel. Results update as you change the inputs.
Choose a preset to load an example, then adjust the values for your situation.
See the inputs and their calculated results.
True Positives (TP): 90 count; False Positives (FP): 10 count; True Negatives (TN): 85 count; False Negatives (FN): 5 count
Specificity: 0.8947368421 percentage; Recall (Sensitivity): 0.9473684211 percentage; Precision: 0.9 percentage; Accuracy: 0.9210526316 percentage; F1 Score: 0.9230769231 percentage
True Positives (TP): 60 count; False Positives (FP): 20 count; True Negatives (TN): 50 count; False Negatives (FN): 10 count
Specificity: 0.7142857143 percentage; Recall (Sensitivity): 0.8571428571 percentage; Precision: 0.75 percentage; Accuracy: 0.7857142857 percentage; F1 Score: 0.8 percentage
True Positives (TP): 70 count; False Positives (FP): 30 count; True Negatives (TN): 40 count; False Negatives (FN): 10 count
Specificity: 0.5714285714 percentage; Recall (Sensitivity): 0.875 percentage; Precision: 0.7 percentage; Accuracy: 0.7333333333 percentage; F1 Score: 0.7777777778 percentage
Use calculators for algebra, geometry, trigonometry, probability, statistics, averages, z-scores, tests, and graphing tasks.
Continue with another tool on a related topic.
Performs a t-test (one-sample, two-sample, or paired) to compare means and determine statistical significance.
Finds the Z-score corresponding to a given cumulative probability (percentile) on a normal distribution.
Calculates the Chi-Square statistic and p-value for goodness-of-fit or test of independence. It requires observed and expected frequencies or values.
Identifies potential outliers in a dataset using the IQR * 1.5 rule (calculating fences).