Abstract—Analyzing and understanding customer requirements is strongly related to the long-term growth of businesses. The lack of automatic methods to elicit customer requirements makes it a time- and money-consuming task for companies to obtain useful information for product design and marketing specialists. In the case of the e-commerce industry, the vast amount of customer feedbacks makes it unfeasible to assess the needs of customers manually. In the presented study, Latent Dirichlet Allocation is applied in combination with Part of Speech selection to extract product and service requirements from a large collection of Japanese reviews. Results obtained in experiments suggest that the proposed approach can be used in practice to efficiently and effectively manage customer requirements in the e-commerce industry.
Index Terms—Requirement elicitation, customer need assessment, topic modeling, Japanese language.
The authors are with Graduate School of Information Science and Engineering, Ritsumeikan University, Japan (e-mail: gr0278ir@ed.ritsumei.ac.jp).
[PDF]
Cite: Mate Kovacs and Victor V. Kryssanov, "A Semi-automatic Approach for Requirement Discovery in the E-commerce Industry," International Journal of Knowledge Engineering vol. 4, no. 1, pp. 68-71, 2018.