Feature usage and access pattern evaluation for set-top-box feature recommender systems

Abstract

In consumer electronic products, such as Set-Top-Boxes (STBs) or Television (TV) user interfaces, usability is a key element which determines the acceptance of new technologies and features by users. The usability analysis questionnaires are limited and misleading. On the other hand, in the big data era, with Internet-connected devices, it is easy to obtain consumers’ usage data from these electronic products. This paper presents a Markov Chain based algorithm to increase the usability of these devices by analysing behavioural patterns of users to support user-centred feature recommender systems. The proposed algorithm identifies user habits, use of device features and access to these features by processing user usage history. The method is evaluated using the real-world data collected from STBs of test users with a broad distribution of age, occupation and technological experience. Results show that the method extracts the behavioural patterns of the users efficiently.