Learning to Rank算法基本原理
Learning to Rank主要有3种算法：
- Pointwise: Subset Ranking, McRank, Prank, OC SVM
- Pairwise: Ranking SVM, RankBoost, RankNet, GBRank, IR SVM, Lambda Rank, LambdaMart
- Listwise: ListNet, ListMLE, AdaRank, SVM MAP, Soft Rank
评价指标
信息检索中常用的有ndcg:
Pointwise算法
逻辑回归交叉熵: 
留下评论