Klusterisasi Fitur Tanya Dokter Pada Website Alodokter Menggunakan Metode Latent Dirichlet Allocation

Abstract

Digital advancements have change information-seeking behaviors, particularly in health inquiries for the people. The Alodokter website's "Tanya Dokter" feature facilitates an easy connections with medical experts to ask question regarding health. The posed questions tend to mirror evolving health trends and public misunderstandings regarding health issues. Manual analysis of data in “Tanya Dokter” features proves challenging, prompting the use of Latent Dirichlet Allocation (LDA) topic modeling. This research categorizes Alodokter topics, unveiling common health issues. The optimal model reveals 8 clusters with diverse topic distributions. Validation metrics using coherence score with 0.258481 as the highest value affirm the model's efficacy. Optimal outcomes stem from combination of parameter such as 8 topics, alpha 0.02, and beta 0.02. This study may offers Alodokter and healthcare providers an informed perspective on an accessible approach to categorize health questions effectively using Topic Modelling  Latent Dirichlet Allocation.