An Integration Module for Mining Textual and Visual Social Media Data

Date

2018-11

Type

Conference paper

Conference title

The Seventh International Conference on Data Analytics (DATA ANALYTICS 2018)

Author(s)

Mohammed Eltaher

Abstract

In social media networks, visual data, such as images, co-exist with text or other modalities of information. In order to benefit from different data modalities, further research is obligatory. This paper introduces the first step towards an integration module, which is based on visual and textual data available in social media. We utilized tags and images of users with a novel approach of information fusion in order to enhance the social user mining. Here, two different approaches were applied to enhance social user mining: (1) content based image fusion, and (2) semantic based image fusion. Our approaches were applied to a gender classification mining application and showed the performance of our methods to discover unknown knowledge about user.