Sentiment Analysis of Ecommerce Product Review Data Based on Deep Learning
Rong Li, Weibai Zhou, Huang Debo
Abstract
Aiming at the problem that traditional sentiment analysis methods are not effective in mining e-commerce product review data, we propose a sentiment analysis model based on deep learning for e-commerce product review data. In our model, comments are divided into positive and negative categories. The text is divided into words, and the word vector is combined with the word frequency to input into the neural network for training. Convolution neural network is used to mine the deep association between feature set and emotion tag, and train emotion classifier. The experimental results show that the model can achieve effective product feature extraction and high accuracy of sentiment classification, which is an effective model for online review analysis.