Latent Dirichlet Allocation (LDA) for Sentiment Analysis Toward Tourism Review in Indonesia

International Conference on Computing and Applied Informatics 2016

物理学;计算机科学

Putri, I.R.^1 ; Kusumaningrum, R.^1

Department of Informatics, Faculty of Science and Mathematics, Diponegoro University, Semarang, Indonesia^1

关键词: Foreign exchange;    Indonesia;    Latent dirichlet allocations;    Negative sentiments;    Probabilistic topic models;    Southeast Asia;    Statistical datas;    Tourism industry;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/801/1/012073/pdf
DOI  :  10.1088/1742-6596/801/1/012073

学科分类：计算机科学（综合）

来源: IOP

【 摘 要 】

The tourism industry is one of foreign exchange sector, which has considerable potential development in Indonesia. Compared to other Southeast Asia countries such as Malaysia with 18 million tourists and Singapore 20 million tourists, Indonesia which is the largest Southeast Asia's country have failed to attract higher tourist numbers compared to its regional peers. Indonesia only managed to attract 8,8 million foreign tourists in 2013, with the value of foreign tourists each year which is likely to decrease. Apart from the infrastructure problems, marketing and managing also form of obstacles for tourism growth. An evaluation and self-analysis should be done by the stakeholder to respond toward this problem and capture opportunities that related to tourism satisfaction from tourists review. Recently, one of technology to answer this problem only relying on the subjective of statistical data which collected by voting or grading from user randomly. So the result is still not to be accountable. Thus, we proposed sentiment analysis with probabilistic topic model using Latent Dirichlet Allocation (LDA) method to be applied for reading general tendency from tourist review into certain topics that can be classified toward positive and negative sentiment.