Trend Topic Analysis using Latent Dirichlet Allocation (LDA) (Study Case: Denpasar People’s Complaints Online Website)

Aulia Rizki Destarani, Isnandar Slamet, Sri Subanti


According to the publication of the Central Bureau of Statistics 2017, the population of Denpasar people has increased to 914,300 people. The Increasing number of the population raises various problems that must be faced by the Denpasar’s Government. The variety of problems is in line with the increase in complaints data posted through Denpasar people’s complaints online website, which made it difficult to know the main topics of the problems. The purpose of this research is to find the main topics of complaints Denpasar residents quickly and efficiently. The method used to achieve the objective of the research is Latent Dirichlet Allocation topic models with Gibbs sampling parameter estimation. The number of topics obtained through the highest log-likelihood value -42,528.84, the value is in the number of topics 19. The trending topic was based on the highest topic probability, topic 4, with a topic probability value 0.055. Based on these results, the trend of a topic is on topic 4 which can be interpreted that many residents of Denpasar complained about damaged roads and requested to fix the roads.


Trend Topic; Topic Models; LDA; Gibbs sampling; Denpasar

Full Text:



D. M. Blei and J. D. Lafferty, “Topic Models,” Text Min. Classif. Clust. Appl., 2009 available at: Google Scholar.

J. Tang, Z. Meng, X. L. Nguyen, Q. Mel, and M. Zhang, “Understanding the limiting factors of topic modeling via posterior contraction analysis,” in 31st International Conference on Machine Learning, ICML 2014, 2014 available at: Google Scholar.

J. Jayadharshini, R. Sivapriya, and S. Abirami, “Trend square: An Android Application for Extracting Twitter Trends Based on Location,” in Proceedings of the 2018 International Conference on Current Trends towards Converging Technologies, ICCTCT 2018, 2018, doi: 10.1109/ICCTCT.2018.8551056

C. Zou, “Analyzing research trends on drug safety using topic modeling,” Expert Opin. Drug Saf., pp. 629-636 2018, doi: 10.1080/14740338.2018.1458838