A Classification of Thai Youth on Using Social Network from Online Behaviours

Abstract

In this paper, the text classification is advanced from a topic modelling perception. We present an application of the Latent Dirichlet Allocation (LDA) topic model: to classify an unstructured text into semantically coherent sectors. The main advantage of the recommended methodology is that along with the sector limitations, it produces the topic distribution associated with each sector. We express an alternate methodology that converts message posting data into a text representation and uses topic modelling to identify outlines through the data set. The data were collected from 1,391,549 comments posted in 15 Facebook fan pages between June 2019 and January 2020. We establish that the behaviour possibility of using social media has become progressively relation term; education and entertainment have come to be more significant. By quantifying the sequential movements of topics at the generation level, we found that Thai youth in diverse areas tend to focus on altered sub-domain, while closely generation normally share educational topics. Our discoveries could be of excessive value to academics and policy makers who are interested in classifying encouraging or popular topics. The results demonstrate topic modelling as a beneficial implement for mining significant data from realistic Thai youth social media using behaviours.