Emotion-Based Content Personalization in Social Networks

Abstract

Personalization is the process of customizing social network pages of users according to their needs and personal interests. It can also be used for filtering unwanted information from an individual's page received from other users, in case this information is unpleasant or unacceptable. To avoid unwanted information from a particular user in current social networks, the user needs to be denied accessibility by blocking them. However, instead of blocking the user, it would be preferable to have a mechanism that prevents the undesirable content in a user's social network page. Thus, this paper presents a model that determine the emotions shared in the content of a social network page by the user. The model determines the dominant emotions for a period of time and uses these to filter the content using the user's dominant emotions. Using the developed model, a novel system based on item based collaborative filtering process to personalize the user's social network page has been developed. A user study involving 5000 Twitter messages shows that the developed system performs satisfactory with a correctness in the filtering process of 87%.