A proactive personalised mobile recommendation system using analytic hierarchy process and Bayesian network

Abstract

With the growth, ready availability and affordability of wireless technologies, proactive context-aware recommendations are a potential solution to overcome the information overload and the common limitations of mobile devices (inconvenience of data input and Internet browsing). The automatic provision of just-in-time information or recommendations tailored to each user's needs/preferences contextualised from the user's activities, location, usage patterns, time, and connectivity may not only facilitate access to information but also remove barriers to the adoption of current and future services on mobile devices. This paper describes a hybrid P2P context-aware framework called JHPeer which supports a variety of context-aware applications in mobile environments. Any context-aware information services such as recommendation services could use the collected and shared contextual information in JHPeer network. An analytic hierarchy process based multi-criteria ranking (AHP-MCR) approach has been developed and used to rate recommendations in a variety of domains. The weights of the contexts criteria can be assigned by the user or automatically adjusted via individual-based and/or group-based assignment. Additionally, a Bayesian network algorithm is applied to solve the cold-start problem in recommendation systems. The paper also proposes a strategy for using Bayesian networks for recommendation services. A news recommendation application has been implemented on the

A Survey on Personalized News Recommendation Technology

“…And some researches [51] show that the use of multi-criteria approach makes it easier to integrate contextual information into the recommendation ranking methods. In addition, AHP-MCR [4] (Analytic Hierarchy Process based Multi-Criteria Ranking) is proposed to deal with the contextual information with dynamic real-time changes during the ranking process. It only gives a general AHP hierarchical model to support the function of adding or deleting contextual criteria or adjusting weights of different criteria flexibly corresponding to different application scenarios or user demands, rather than designing different AHP models for every different applications.…”

“…W is the weights corresponding to the terms. The commonly used methods of term weights calculation in VSM include frequency statistics [57], TF-IDF (Term Frequency-Inverse Document Frequency) [4] and a series of improved methods, among which TF-IDF method is the most widely used. With the development of text mining technology, the structured representation of news texts is also getting more and more mature, which is conducive to the further development of personalized news recommendation.…”

“…And then with the guide of latent ''experts'', the cold start problem may be solved. Yeung et al [4] propose a strategy using Bayesian network to eliminate the data sparseness problem for new users, that is, the user cold start problem. A Bayesian network is a probabilistic graphical model which combines the advantages of CF and CB.…”

“…And three phases (i.e. construction, prediction and revision) are designed to resolve the cold start problem in [4]. A hybrid recommendation method is used in [19], which takes into account many recommendation factors such as news freshness, news popularity, users' long-term and short-term interests and users' contextual information.…”

“…On the one hand, there are countless sources of news and massive news, making it difficult for users to make choices among vast amounts of news. On the other hand, different news websites and apps have different resources and backgrounds, leading to messy contents of news [4]. Personalized news recommendation technology [46] applies personalized recommendation technology in the field of news.…”