SYSTEM TO FILTER UNWANTED MESSAGES FROM OSN USER WALLS

Abstract

One fundamental issue in today’s Online Social Networks (OSNs) is the lack of effective mechanisms enabling users to control the messages posted on their private spaces, such as walls, to prevent the display of unwanted or inappropriate content. Currently, most OSNs offer minimal support for users to filter or moderate such content, leaving them exposed to spam, offensive messages, or irrelevant posts. To address this gap, this paper introduces a novel system that empowers OSN users to directly manage and regulate the messages appearing on their walls. The proposed system leverages a flexible rule-based framework that allows users to define and customize filtering criteria according to their personal preferences and privacy requirements. In addition, the system integrates a Machine Learning-based soft classifier capable of automatically labelling messages based on their content, facilitating a more effective content-based filtering process. This combination of rule-based customization and automated classification provides a robust and user-friendly solution to enhance user control over their online interactions. Experimental results demonstrate the system’s ability to accurately filter undesired content, significantly improving the user experience and safety in online social environments.