Extracting Product Features And Determine Sentiment Orientation From Chinese Online Reviews

With the rapid development of the Internet and E-commerce, Internet is becoming an important platform for consumer to write product reviews, product reviews are growing at an alarming rate. These reviews contain a lot of information, manufacturers and users can effectively use this information to make decisions. By browsing the reviews of a product on the Internet, users can learn about users’ attitudes distribution tend to a product, then make purchasing decisions; manufacturers can understand users’feedback on its products, also evaluation of their own and competitors, thereby improving the product by users’feedback to gain competitive advantage.But the product reviews on the Internet are a huge number of unstructured data, if the manufacturers and users read reviews through artificial way to get the information they need, not only time consuming and prone to errors. Consequently, manufacturers and users urgently need the help of some technical means to process a large number of product reviews automatically, which can enable them to quickly get the accurate information they need. It is because of the urgent needs of manufacturers and users, reviews mining has become a hot topic. Reviews mining use association rules, semantic analysis, machine learning and statistical analysis techniques to process a large number of reviews that users write after they buy it, and then summed up users’ attitude tend distribution to the product features automatically. Users can make purchase decisions based on mining results quickly and accurately; manufacturers can get the features to be improved, also the product features to be propaganda according to mining results, this can enhance the product’s response capabilities to the market and improve users’satisfaction.Based on semantic and linguistic analysis, use the semantic relationships between product features and opinion words and the polarity scores of opinion words influence on product features. This paper proposes a method which can extract product features from the online reviews and rank them based on customer attention strength. At the same time, the sentiment orientation users have on the feature is determined according to the polarity scores of opinion words. The online reviews about notebook Computers are used as experimental data. Experimental results show that the method achieved good precision and recall rate.