In this paper, we present our work on analyzing the impact of the Mudflow Disaster in Sidoardjo, Indonesia, based on text mining technologies. We conducted a topic extraction using the Latent Dirichlet Allocation model. To handle the difficult expressions and grasp the points, we use various techniques such as bigram segmentation for documents related to the Mudflow in English. The TreeTagger is the morphological analysis tool used. The extracted topics clearly showed the impact of the Sidoardjo Mudflow. The most widely discussed topic found was the resettlement conditions and the compensation for the victim corresponding to the presidential regulation. We also found other frequently mentioned topics, such as the payment of resettlement, water pollution, and the verification process for the households.
|Journal||Gakushuin University Economics|
|Publication status||Published - 1 Oct 2016|
- Topic extraction, Dirichlet Allocation Model, Sidoardjo Mudflow, Compensation, Resettlement, Presidential Regulation.