Ordinal Regression And Ranking
DEMAND-DRIVEN ASSOCIATIVE CLASSIFICATION(2011)
摘要
Accurate ordering or ranking over instances is of paramount importance for several applications (Faria et al. Learning to rank for content-based image retrieval. In: Proceedings of the Multimedia Information Retrieval Conference, pp. 285-294, 2010; Veloso et al. Learning to rank at query-time using association rules. In: Proceedings of the Conference on Research and Development in Information Retrieval (SIGIR), pp. 267-274, 2008; Veloso et al. J Inf Data Manag 1(3): 567-582, 2010; Veloso and Meira, Efficient on-demand opinion mining. In: Proceedings of the Brazilian Symposium on Databases (SBBD), pp. 332-346, 2007; Veloso et al. Automatic moderation of comments in a large on-line journalistic environment. In: Proceedings of the International AAAI Conference on Web logs and Social Media (ICWSM), pp. 234-237, AAAI, 2007).One clear application is Information Retrieval, where documents retrieved by search engines must be ranked according to the corresponding relevance to the query (Trotman, Inf Ret 8(3):359-381, 2005).Many features may affect the relevance of such documents, and, thus, it is difficult to adapt ranking functions manually. Recently, a body of empirical evidence has emerged suggesting that methods that automatically learn ranking functions offer substantial improvements in enough situations to be regarded as a relevant advance for applications that depend on ranking. Hence, learning ranking functions has attracted significant interest from the machine learning community. In the context of Information Retrieval, the conventional approach to this learning task is to assume the availability of examples (i.e., a training data, S, which typically consists of document features and the corresponding relevance to specific queries), from which a learning function can be learned. When a new query is given, the documents associated with this query are ranked according to the learned function (i.e., this function gives a score to a document indicating its relevance with regard to the query). In this chapter we present ranking algorithms based on demand-driven associative classification.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络