Personalised Dynamic Viewer Profiling for Streamed Data

Abstract

Nowadays, not only the number of multimedia resources available is increasing exponentially, but also the crowd-sourced feedback volunteered by viewers generates huge volumes of ratings, likes, shares and posts/reviews. Since the data size involved surpasses human filtering and searching capabilities, there is the need to create and maintain the profiles of viewers and resources to develop recommendation systems to match viewers with resources. In this paper, we propose a personalised viewer profiling technique which creates individual viewer models dynamically. This technique is based on a novel incremental learning algorithm designed for stream data. The results show that our approach outperforms previous approaches, reducing substantially the prediction errors and, thus, increasing the accuracy of the recommendations.

Data-driven personalisation of television content: a survey

“…In [197], the profiling of TV viewers is addressed in a different way making use implicitly of user feedback to online content. The paper proposes a personalised viewer profiling technique that creates individual viewer models dynamically using an incremental learning algorithm to learn from viewer comments, likes and shares on streamed content.…”

“…TV personalisation and interaction: One of the biggest service sectors enabled by object-based media is for personalisation of and interaction with TV [221,222]. In these scenarios, a user may have content personalised for them on the basis of a number of factors, for example, a user profile, previous interactions with the media, etc.…”