Recall score, true positive rate검색된 정보 중에서 관련 있는 정보가 얼마나 잘 검색되었는지Sensitivity=P(Predict=1∣Actual=1)Sensitivity = P(Predict=1|Actual=1)Sensitivity=P(Predict=1∣Actual=1)TPTP+FN\frac{TP}{TP +FN}TP+FNTPTrue Positive / True Positive + False Negative Precision and Sensitivity are in a Trade-off relationship, so we use the F1-score. We use it to find a suitable point where both are high.