Tourist Place Classification Using Topic Modeling Algorithms

• Published in: Proceedings - International Carnahan Conference on Security Technology pp. 1 - 10
• Main Authors: Wadhe, Apeksha Arun, Suratkar, Shraddha S.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 11.10.2023
• Subjects: Clustering algorithms; Coherence; Coherence Score; Gibbs Sampling; Inference algorithms; Latent Allocation Dirichlet; Perplexity; Prediction algorithms; Predictive models; Social networking (online); Training; Variational Inference
• ISSN: 2153-0742
• DOI: 10.1109/ICCST59048.2023.10474232

Today social media is one of the most powerful and popular tool which is expanding rapidly in the world. Many tourism websites publish tourist places data for describing places in the form of Description. Over this data, Analysis can be performed and meaningful insights can be drawn. By analyzing this data we can find the subject or topic discussed in Description. By performing topic modeling over these data we can predict the category of tourist place. Also topic Modeling can help to cluster textual content into groups as per the subject of the text. In research study, Comparative analysis of topic modeling algorithms has been performed which will find a category of tourist places. Tourist place description blog data was gathered from tourism websites. Using tourist place Description dataset, comparative study of Topic modeling algorithms i.e. LDA Variational Inference(VI) algorithm and LDA Gibbs Sampling(GS) algorithm has been performed. Results of experiments has been analyzed with performance evaluation using various factors such as coherence score, perplexity, training execution time and accuracy percentage. LDA Gibbs Sampling(GS) algorithm outperforms LDA Variational Inference (VI) algorithm in terms of coherence score, perplexity, accuracy percentage. Whereas, LDA Variational Inference(VI) algorithm requires less execution time as compared to the LDA Gibbs Sampling(GS) algorithm