Sentiment classification method of commodity comment data based on convolutional neutral network

Abstract

The invention discloses a sentiment classification method of commodity comment data based on a convolutional neutral network. Focusing on the problem that according to traditional word vector representation methods, only vocabulary characteristics and syntax characteristics are extracted, while the semantic relationships among words are ignored, and aiming at the purpose that the classification effects of traditional classification algorithms can be improved, the sentiment classification method of the commodity comment data based on the convolutional neutral network is put forward. Initially the word vector characteristics of each piece of comment data are extracted by means of a CBOW model, the semantic relationships among the words and the representation capability of the word vector characteristics are enhanced, then a sentiment classification model is built by means of the deep learning network model, the convolutional neutral network model, and the sentiment classification performance of the comment data is improved. According to the technical scheme, the method has the advantages of being simple and fast, and the sentiment classification performance of the comment data can be well improved.

Description

technical field [0001] The invention belongs to the technical field of sentiment classification, in particular to a method for sentiment classification of commodity review data based on a convolutional neural network. Background technique [0002] (1) Sentiment classification technology [0003] With the rapid development of mobile network technology and the popularity of smart phones, people tend to buy goods directly on some e-commerce websites through mobile APPs, and users will also post comments on these e-commerce websites to share their purchases. The feeling of using the product after purchasing the product. Sentiment analysis on commodity review data, also called opinion mining, is the process of collecting, preprocessing, and discriminating the emotional tendency expressed by customers on commodity review information. Sentiment classification of review data on e-commerce websites can determine the buyer's emotional tendency for a product: whether he likes the pro...

AI Technical Summary

Problems solved by technology

[0016] Aiming at the problem that the traditional word vector representation method only extracts lexical features and syntactic features, but ignores the semantic relationship between words, the traditional classification algorithm classification effect can be further improved, and provides a convolution-based Sentiment classification method for commodity review data based on neural network

Images