Increasing Serendipity of Recommender System with Ranking Topic Model

Abstract

There are thousands of academic paper published each year, it is quite hard for researchers who enters a new field to discover relevant paper and novel paper to read, which we characterize as choice overload problem. Recommender system can help to alleviate the problem, but recommender system suffers from the intention gap problem which is the incapability of the system to accurately guess users' intentions. We proposed a ranking topic model based semantic recommendation framework which helps to introduce serendipity to the system. First, the proposed ranking topic model reorders learnt topic distributions according to users' intentions. Then, learnt ordered topics are used as features to rank papers in the library according to the relevancy to user query. At the same time, ranked topics also provide novelty to the results. Since there is little work on how to evaluate the serendipity degree of recommender system, we proposed two measure to evaluate this metric. We performed empirical experiments to test the efficiency of proposed framework with state-of-the-art counterparts, the comparison results revealed the superiority of our proposed algorithms. In the end, we illustrated our algorithms with an example and pointed out future research directions.

Relevance, diversity and serendipity in content recommendation using clustering

“…Estratégias para embutir e avaliar surpresa e serendipidade em recomendações são estudadas em [Zheng and Ip 2012, Jenders et al 2015, Piao and Whittle 2011, Xiao et al 2014. Em [Zheng and Ip 2012], um framework foi desenvolvido para equilibrar os graus de surpresa de recomendações.…”

“…Em [Piao and Whittle 2011], com uso de processamento de língua natural (NLP), os autores extraem os interesses dos usuários e sugerem conexões serendipitosas. Em [Xiao et al 2014], os autores propõem um framework de recomendações serendipitosas de artigos acadêmicos e duas medidas que avaliam tal característica.…”

Graph-based collaborative ranking

“…The quality of recommendation can be analyzed from many different points of view including accuracy (Koren, Bell, & Volinsky, 2009;Weimer & Karatzoglou, 2007), coverage (Bellogin & Parapar, 2012;Cacheda, Carneiro, Fernández, & Formoso, 2011), diversity (Adomavicius & Kwon, 2012;Said, Kille, Jain, & Albayrak, 2012;Zhou et al, 2010), serendipity (Lu, Chen, Zhang, Yang, & Yu, 2012;Xiao, Che, Miao, & Lu, 2014), uncertainty (M. Zhang, Guo, & Chen, 2015), shilling attack detection(Z. Zhang & Kulkarni, 2014) and scalability (Jiang, Lu, Zhang, & Long, 2011).…”

Opinion-enhanced collaborative filtering for recommender systems through sentiment analysis

“…A promising solution is to cluster the raw data (Liu et al, 2013;Xu, Zheng, & Ding, 2012) or employ the topic model (Xiao et al, 2014). A sophisticated alternative is a multi-criteria CF method, also called as hybrid recommendation (Nilashi et al, 2014).…”