Contextual Recommender Systems Using a Multidimensional Approach

Abstract

Abstract-Recommender systems use the past experiences and preferences of the target users as a basis to provide personalized recommendations for them and as the same time, solve the information overloading problem. Context as the dynamic information describing the situation of items and users and affecting the user's decision process is essential to be used by recommender systems. Multidimensional approach to recommender systems that can provide recommendations based on additional contextual information besides the typical information on users and items used in most of the current recommender systems. This approach supports multiple dimensions, profiling information, and hierarchical aggregation of recommendations. the recommender system could simultaneously possess the advantages of content-based recommendation, knowledge-based recommendation, collaborative filtering recommendation and OLAP in segmenting the information. Following the improvement of the recommendation structure, it doesn't have to limit its analysis on the user and product to compute for the recommendation result and it could also handle and determine more complex contextual information as recommendation computation foundation. It could develop better results if applied in different domains.

An Approach for Integrating Multidimensional Database into Context-Aware Recommender System

“…[15] In Table 2 and Table 3, we summarize criteria being attributes of user dimension and Destination dimension. [6], [8], [11], [13], [14] Yes. This attribute highly effect to user 's decision mentioned in Table 1 Destination Category [6], [13], [15], [16] Yes Region [15] Yes Activity [6], [13], [15], [16]…”

“…There are many researches on the CONTEXT definitions and one of general definitions is "conditions or circumstances which affect something" [1] [4]. In personalized Recommender Systems, context information has been recognized as an important factor to be considered [1] [6]. Therefore, Adomavicius et al in [1] propose a multidimensional approach for presenting context information beside the information of users and items in RS in order to increase the quality and performance of recommendation.…”

“…For example, we may want to know how individual user "Mr.A" like a destination place "Nha Trang City", we need to compute R(Mr.A, Nha Trang) that is the predicted rating of "Mr.A" to "Nha Trang". Because beach is a feature of the "Nha Trang" destination, we can use the hierarchy that "beach"is higher level in destination place dimension to compute aggregate rating R ( In [6], the authors propose a rating estimation function for user u to item i in a certain context c referred as r(u, i, c). That function in [6] is based on: (1) historical ratings of user u to all items being similar to item i referred as aggregate computation R(u, rollup i, c) like content-based concept or (2) historical rating of all users being similar to user u to item i referred as aggregate computation R(rollup u, i, c) like collaborative filtering concept.…”

Context-Based Multi-Agent Recommender System, Supported on IoT, for Guiding the Occupants of a Building in Case of a Fire

“…For Jannach et al [21], exploring the user's location-knowing who accompanies them and what nearby resources are available-increases the quality of recommendations. Rahman [22] also mentions that taking the context into account contributes to the improvement and reliability of recommendations. Thus, because they consider the contextual situation of the user, context-based recommender systems can generate more relevant recommendations [23].…”