Date of Publication :14th June 2017

Abstract: As a symptom of progressively prevalent online networking, cyberbullying has developed as a difficult issue afflicting kids, youths and youthful grown-ups. Machine learning methods make programmed identification of harassing messages in web-based social networking conceivable, and this could develop a sound and safe web-based social networking condition. In this significant research region, one basic issue is vigorous and discriminative numerical portrayal learning of instant messages. In this paper, we propose another portrayal learning technique to handle this issue. Our strategy named Semantic-Enhanced Marginalized Denoising Auto-Encoder (smSDA) is produced through semantic expansion of the prominent profound learning model stacked denoising autoencoder. The semantic expansion comprises of semantic dropout commotion and sparsity requirements, where the semantic dropout clamor is outlined in light of area learning and the word inserting system. Our proposed strategy can misuse the concealed element structure of tormenting data and take in a hearty and discriminative portrayal of content. Far reaching investigates two open cyberbullying corpora (Twitter and MySpace) are directed, and the outcomes demonstrate that our proposed approaches outflank other standard content portrayal learning strategies.

Reference :