Offline Evaluation Metrics

Description of this Post

1 Outline

Slide 25

2 Precision and recall

Slide 26

3 Precision and recall

Slide 27

4 Precision-recall curve

Slide 28

5 F-measure

Slide 29

6 Any problems with the metrics so far?

Slide 30

7 Outline

Slide 31

8 Precision and recall

Slide 32

9 Other common metrics

Slide 33

10 Any problems with the metrics so far?

Slide 34

11 Outline

Slide 35

12 User search behavior

Slide 36

13 Discounted cumulative gain (DCG)

Slide 37

14 Rank-biased precision (RBP)

Slide 38

15 Rank-biased precision (RBP)

Slide 39

16 Expected reciprocal rank (ERR)

Slide 40

17 Expected reciprocal rank (ERR)

Slide 41

18 Offline evaluation summary

Slide 42

19 Materials

Slide 43

20 Materials

Slide 44