Text-Based Intelligent Content Filtering on Social Platforms

Abstract

Social platforms have become one of the popular mediums of information sharing and communication over the Internet today. People share all types of contents such as text, images, audio and video using these social platforms. Though information gained using these social platforms can be very useful for people around the globe, some of the user generated contents are very negative as they contain abusive, racial, offensive and insulting material. Thus, there is a need for an effective online content filtering technique which blocks these negative contents while not disturbing the access of users to rest of the contents available on these sites. Current techniques simply filter on the basis of URLs blocking and keyword matching or either rely on a large database of pre-classified web addresses. The problem is how to intelligently filter the negative contents, rather than filtering entire websites using their URLs or applying simple keyword matching techniques. In this paper we review a number of existing approaches to content filtering and propose an intelligent content filtering technique that uses sentiment analysis of the text and feature engineering methods to perform text classification.

Web Content Filtration Using Different Web Mining Techniques in Educational System: An Overview

“…Another J48 decision tree classifier is used to find relationship between features and classes. [1] In this paper Jianping Zhang, Jason Qin and Quiuming Yan, have done work on a novel URL based objectionable content categorization approach and application of web filtering. In this model Maximum entropy algorithm, machine learning algorithms are used to break URL in n-grams.…”

Text Filtering through Multi-Pattern Matching: A Case Study of Wu–Manber–Uy on the Language of Uyghur

“…Unfortunately, more and more often, big data and big information are mixed with erroneous and mistrustful noises. A booming growth in social media in recent years in particular, calls for development and progress in techniques and tools for filtering out textual noises [1]. Regarding that aspect, content-based filters and models are particularly successful [2], and the ones for the English language have already been widely used in various fields, including information retrieval [3,4], network security [5,6], personalized information extraction and inference [7], content-based recommendation systems [8], knowledge discovery [9,10], Short Message Service (SMS) spam filtering [11,12], and many other areas.…”