Rank Based Metric
간단히 생각하려면 relevance score와 position score를 계산하는 방식들의 차이라고 요약할 수 있다. mrr은 position의 역수를 position score로 두고 relevance score는 전혀 고려하지 않는 방식이라면 DCG는 relevance score를 log2(position score)로 나누어준 값을 사용한다.
commonly used
1.Mean Reciprocal Rank (MRR)
2.Normalized Discounted Cumulative Gain (nDCG) - It measures the gain of the retrieved items relative to the ideal ranking of these items.
3.Precision@k - It measures the proportion of relevant items among the first k retrieved items.
4.Mean Average Precision (MAP) - It is the average of the precision scores at each relevant item position in the ranking.
5.F1-score@k - It is the harmonic mean of precision and recall, computed for the first k retrieved items.
6.Receiver Operating Characteristic (ROC) curve - It is a graphical representation of the performance of a binary classifier.
7.Area Under the ROC Curve (AUC) - It is the area under the ROC curve, which summarizes the performance of a binary classifier across all possible thresholds.
8.Kendall's Tau rank correlation coefficient - It measures the strength of the association between two variables and its value is between -1 and 1.
reference site
https://lamttic.github.io/2020/03/20/01.html
https://medium.com/swlh/rank-aware-recsys-evaluation-metrics-5191bba16832
