Topic prediction modelling on social media content using machine learning

Izmi Dewi Aisha, Lili Ayu Wulandhari


The simplicity to deliver an opinion about companies or institutions via social media has resulted in both positive and negative judgments. Through social media all positive and negative information will be easily found and spread. It is concerned that negative information will lead to negative public opinion. If this occurs, the company will suffer from a lack of trust, which will harm the company's reputation. Thus, to monitor uncontrolled issues, a company wants to know what topics or opinions are developing in the community. Therefore, the topic modelling using latent dirichlet allocation (LDA) is proposed to identify topics that are being discussed on social media. The findings of this study got the coherence score of 0.558 and based on the direct human judgment, the model got an average 80% correctly. The findings of this study reveal 4 topics groups that represent the corporate social media content. These findings offer information to companies about the latest topics or opinions that are currently developing in society which could provide recommendations related to decision-making on current issues thus increasing the trust and reliability towards the company.


Clustering; Corporate social media; Latent dirichlet allocation; Topic modelling; Topic prediction

Full Text:




  • There are currently no refbacks.

Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

The Indonesian Journal of Electrical Engineering and Computer Science (IJEECS)
p-ISSN: 2502-4752, e-ISSN: 2502-4760
This journal is published by the Institute of Advanced Engineering and Science (IAES) in collaboration with Intelektual Pustaka Media Utama (IPMU).

shopify stats IJEECS visitor statistics