Thai Variable-Length Question Classification for E-Commerce Platform Using Machine Learning with Topic Modeling Feature

Abstract

At present, online shopping is a part of our life. Either a new joiner or an expertise sometimes has questions regarding applications. The most convenient and effective way is to contact the customer service via live chat. However, a huge number of customers causes a long waiting time affecting customers' experience. Thus, this article proposes Thai variable-length question classification for e-commerce platform to deal with this problem. A fusion of two model architectures, Latent Dirichlet Allocation (LDA) and Long Short-Term Memory (LSTM) has been proposed and used as a feature extraction before applying the softmax function to classify the questions. The experimental results have been shown that the proposed model is able to achieve an accuracy of 84.43% which is better than the other models.