Research On Product Network Evaluation Information Mining Based On Text Processing Technology

More and more users of online consumer start browsing online reviews to learn a lot of word of mouth of ne twork products and services,which helps you make informed decisions. W hile the network custom er reviews as a feedback mechanism also help the service providers improve their services in order to gain a com petitive edge. However, with the e-business arising,the number of reviews is growing rapidly and the content is more complicated, it is very difficult to r etrieve useful knowledge from customers’ reviews,especially difficult to get people’ s perspectives and attitudes from many characters, events, products in a short period. Therefore, it needs technical m ethods to im prove the accuracy and convenience of mining information. Comments are m ainly related to m ining sentiment analysis, feature mining, subjective content identification; where in emotional orientation analysis object is achieve d by mining and analyzing texts stand, viewpoints, emotions, likes and dislikes and other subjective inform ation, the text of the subjectiv e attitude to judge, involving artificial in telligence, machine learning, data m ining, natural language processing and other fields. In English commentary research, resear chers have m ade some preliminary results, but for the Chinese Internet users-reviewed research still in its infancy. With the ris e of Chinese e-comm erce in the w orld in th e field of Chinese comments about the need to autom atically extract useful information advanced technology. In this text, the inform ation of Chinese Internet travel booking decisions in a very im portant hotel reviews for the study started to explore. Hotel online reviews are very representative, and is dif ferent from other online reviews by custom ers of its m ore dependent on whether the customer product reservation or purchase play a decisive role; it is the customer’s perception of quality of hotel se rvices academia its use has a hotel service quality research relevant results, but more preclude the use of content analysis, we can not comment on batch processing, application of the results greatly restricted.Based on the above issue, we use m achine learning m ethods for the analysis of the em otional tendency network comment text, designed to provide customers and Chinese com panies in the f ield to p rovide more convenient and scientific review m ining tools. In this paper, an open source framework reptiles get custom er reviews evaluated from Corpus Pa cific Internet according to five dim ensions into the categories; Key has details of preconditioning corpus, incl uding the Chinese word and go without words; then choose the feature extraction Na ive Bayes classification m ethods and standards, combined with customer reviews emotion model proposed in the locale multiple algorithms to achieve a f urther improvement of the classification results; results show that th is route is calculated to obtain better classification results, higher accuracy, not only to overcome noise problems in the text analysis of high-dim ensional sparse data problem and foc us on training, and has a stable m ass text segmentation for practical performance, results also show that the analys is of this tendency is more accurately classified and accurately reflect cu stomer the views and position s to help managers quickly grasp customer favorite or aversion for all aspects of mobile phone products, which has practical significance.