stats

• Univariate statistics: analyzing one predictor at a time.
• Multivariate statistics: analyzing several variables together.
• Correlation and regression: measuring relationships and controlling for other variables.

False positives and false negatives belong more specifically to classification evaluation or diagnostic testing. They come from a confusion matrix:

                     Actually positive      Actually negative

━━━━━━━━━━━━━━━━━━━━ ━━━━━━━━━━━━━━━━━━━━━ ━━━━━━━━━━━━━━━━━━━━━ Predicted positive True positive (TP) False positive (FP) ──────────────────── ───────────────────── ───────────────────── Predicted negative False negative (FN) True negative (TN)

Common divisions you may be remembering:

• Accuracy = (TP + TN) / all cases How often was the prediction correct overall?

• Precision = TP / (TP + FP) Of the positive predictions, how many were right?

• Recall / sensitivity / true-positive rate = TP / (TP + FN) Of the actual positives, how many did we find?

• Specificity = TN / (TN + FP) Of the actual negatives, how many did we reject correctly?

• False-positive rate = FP / (FP + TN)

• False-negative rate = FN / (FN + TP)

• Precision asks: “Of the queries we flagged as expensive, how many really were?”

• Recall/hit rate asks: “Of all truly expensive queries, how many did we flag?”