Abstract
|
In earlier times, there are many studies to extract from customers¡¯ product reviews on social networks, e-commerce websites, etc. However, in the Vietnamese market with different characteristics of culture and language have only a few studies published. Thus, in this paper, we present a deep convolutional neural network (D-CNN) architecture using features and linguistic patterns of the Vietnamese language for the aspect extraction task. The archived results are acceptable with F1 scores for three domain book, female fashion, and mobile accessories are 76.92%, 71.74%, and 74.54%, respectively.
|