Building demonstration systems for exploring market competition over topics in spatio-temporal document collections

• Main Author: Kurniawan, Timothy Gabriel
• Other Authors: Cong Gao
• Format: Final Year Project
• Language: English
• Published: 2019
• Subjects: DRNTU::Engineering::Computer science and engineering
• Online Access: http://hdl.handle.net/10356/77085
• Institution: Nanyang Technological University

There is a large amount of spatio-temporal documents were generated from social networks in recent years. These data may help the business owner to assess their business based on popular topics of their brands compared with other brands, since the data may indicate what users think over the products or brands. In this project, we aim to mine the topics and the market competition from spatio-temporal documents within a certain geographical region and time period, and then visualize the result as a web application for demonstration purpose. The mining can be done using the EMart framework [1], which combines CompLDA model for mining topics and Octree-based data structure to divide and combine the spatio-temporal documents to three-dimensional cells.