The Product Reviews Information Of Chinese Sentiment Analysis

With the development of Internet technology, we are gradually entering the era of big data, more and more people are accustomed to post their own thoughts and feelings online. Especially, the rising of online shopping brings a flood of commentary information. To analyze and mining these comment that reflect consumer’s emotional information has become an important means for enterprises to improve products and enhance the competitiveness. There is no doubt that good users’ comments are an intangible asset or no-cost advertising. However, unsatisfactory users’reviews will greatly reduce business reputation. Therefore, enterprises need to collect consumers’product reviews and analyze information timely, in order to discover inadequacies of the product. And take effective measures to improve the quality of goods, to form a good evaluation. However, with the rapid growth of online shopping reviews, to collect, process and analyze them by artificial means is difficult. Most of this information is presented by unstructured or semi-structured form, which increases the difficulty of the analysis. Nowadays, the ideal approach is to choose sentiment analysis technology that belongs to the field of text mining to solve such problems.This article is based on sentiment analysis technology which has become the hot spot. Taking HUAWEI glory phone which is the typical Internet marketing products as an example, firstly, use python software to crawl this useful information from the mass comments. Secondly, process and analyze this information, in order to help enterprises get the real needs of consumers and find the inadequacies of products. Thirdly, design a more user-friendly product to win the favor of customers and obtain industry competitiveness.The main study of this article is divided into four parts. Firstly, use python to get sample data; Secondly, based on JIEBA segmentation algorithm and Harbin Institute of Technology platform (LTP) crawl comments data, then cut into sentence, segmentation, POS Tagging, Dependency Parsing and so on. Thirdly, compare POS Tagging and Dependency Parsing Method in feature words and emotion words extracting aspect to find which is better. Lastly, the innovation of this article is based on the different sentence to build a multi-strategy sentiment analysis method; Meanwhile, the article will give the different weight about the first comments and second comments when calculate the emotional value; What’s more,pay attention to the negative sentences’s processing. The results research these aspects need to be strengthened and improved such as battery, standby time, feeling and game. Especially, the battery and standby time make customers significant dissatisfaction, which indicate producers need to pay sufficient attention. The good news is that the product’s price, camera, pixels, etc. are praised by consumers widely.