Design And Implementation Of Massive Shopping Guide System Based On Comments

In the 21 st century,online shopping has been integrated into our daily life.For most people,“the comments from others who bought the goods” will influent deeply for our decision-making.When the Internet is not yet universal,we can obtain the evaluation of information less and single channel.It is more difficult to get more information for businesses and customers.Meanwhile,the wide application of Internet and e-commerce sites to make our access to information and the number of geometric quantity increasing,we get large amount of information through the different e-commerce sites.From the perspective of the analysis of the evaluation for its product information,most of the sites only summarize scores and other quantitative information,not to do word-processing.While reading text comments is very time-consuming,it is necessary to multiple web sites for more information and summarized them.In addition,because of the limitation of the screen size on the mobile terminal,to read text comments information also is not convenient.This article picked all kinds of comments data on major e-commerce sites,designed intelligent e-shopping system based the mass product reviews data on the mobile terminal.First,to learn the experiences on current review data acquisition,realized enormous quality of information which based the improved K-means clustering algorithm.Through learning prediction technique based on the contents of user preferences of CRF of word segmentation,to realize the analysis of the comments in Chinese information processing.According to the similarity of item name,to generate the suggestion list,for the user to score prediction calculation,by grading forecast results list in reverse order generated the final recommendation.Among the comment object data to retrieve the inverted index structure,the data submitted by the user query requests in advance through the front end index server to complete the word segmentation processing,then the results of processing data using inverted index for the query.Stored by the combined HBase and Hadoop platform to realize the big data distributed storage solution to the problem of big data computation and storage.The experiments prove that this system has realized successfully processed must comment number level,to achieve rapid response to the user's query retrieval demand,to provide users with all kinds of valuable evaluation information about commodities comments.