Article Preview
Top1. Introduction
The increasing amount of information on the web and the emergence of e-commerce have led to the problem of information overload. Because of this problem, it becomes difficult for users to search for items of their interest. Therefore, a recommender system is essential in order to identify items based on user’s interest. The system is an information filtering system that recommends relevant items to users by analyzing the users explicitly mentioned preferences and interests (Ojokoh et al., 2013; Khusro, Ali & Ullah, 2016). It saves a lot of time and effort of users typically involve in issuing different queries about the items of interest, by simply prioritising and personalising large volume of information available at its disposal to find the unknown relevant items needed by the users. This prompted many research efforts on recommender systems (Goldberg et al., 1992; Burke, 2002; Linden, Smith & York, 2003; Miller et al., 2003; Pazzani & Billsus, 2007; Cacheda et al., 2011). Among these systems include Collaborative Filtering (CF) which is the most popular and successful system that provides recommendations to users because it recommends any type of items to users such as books, movies, news, music, web pages and so forth (Chen et al., 2011; Ping & Ming, 2012; Omisore, 2014).
The CF uses a similarity function to recommends items by considering users’ ratings of an item to find the match of rating patterns of some items involving other users with similar interests (Montaner, Lopez & Rosa, 2003; Chen et al., 2011). The similarity function is the core component of the recommendation process because the accuracy defends on the type of similarity function used. Several similarity functions have been developed to predict the correct item rating based on the users preferences/ratings (Sarwar et al., 2001; Lee, Park & Park, 2007; Shen & Zhou, 2010; Weijie, Jing & Liang, 2012; Yang, Ali & Li, 2013; Zhao, Niu & Chen, 2013; Zhu et al., 2014; Latha & Nadarajan, 2015; Yang & Wang, 2016; Fan, Yu & Huang, 2018). Among these functions, the developed function in (Fan, Yu & Huang, 2018) uses a similarity function to punish popular items. However, the function may predict inaccurate interested popular items to the active users due to the penalty function employed by the similarity that highly penalises high popular items. In addition, it also leads to the return of few similar items neighbours because it utilises items with common ratings which decreases the tendency of selecting less popular items.
In this paper, a popularised similarity function (pop_sim) is proposed to provide effective recommendations to users. The pop_sim function introduces a modified punishment function to minimise the penalty on high popular items. The function also employs a popularity constraint which uses the items ratings threshold to increase the chances of choosing less popular items so as to obtain more similar items neighbours. Experimental studies conducted indicate that compared to the existing functions, the proposed pop_sim performs better in terms of reducing MAE and RMSE. Thus, improve the accuracy of the rating prediction.