Enhancing topic clustering for Arabic security news based on k‐means and topic modelling

• DOI: 10.1049/ntw2.12017
• Publication Date: 2021-03-09T12:12:47Z
• Authors: Adel R. Alharbi; Mohammad Hijji; Amer Aljaedi
• Keywords: Document Clustering

Abstract The internet has become one of the main sources news spread as it unleashed information dissemination space, where websites express opinions on entities while also reporting recent or unusual security risks. Recently, many research studies have focused sentimental reflection views and impressions people utilising natural language processing analytical linguistics. Therefore, we collected corpus from popular Arabic that publish articles related to issues, provide light weight preprocessing techniques data is term matrix transformed. We present an intensive lexical‐driven analysis with visualised views, our topic modelling technique can effectively extract significant topics all text different websites. Our experiments validate k‐means clustering algorithm without latent Dirichlet allocation method, adopted various validation measure internally externally. As shown in experiments' results, proposed combined method a high round index rate 87.2%, large number clusters.