| With the rapid development of e-commerce,mall diversity,continuous refinement of commodity types,more and more convenient logistics,consumer online shopping more and more frequent,people go shopping from the traditional physical store consumption patterns to online shopping.Consumers buy and use products,the product rating and comment on the network to share their own experience.These comments are increasing,showing the trend of information overload,while the review contains a lot of information,but also reflects the degree of consumer preferences for the product.More and more people like to make a purchase decision before the first online reference product reviews.A user’s rating usually includes a numerical score and a textual comment that reflects the user’s preference for the different characteristics of the product and the user’s emotional tendencies.Because of the amount of information that is available,it is difficult to read all of them to help make decisions.User comments have also become a hot topic in recent years.In the field of e-commerce,recommendation system has been widely used.In the major electronic business platform interface can be seen "guess you like." Recommendation system to provide personalized recommendations to help people overcome the problem of information overload,the core is through the personalized algorithm,the use of different users for different products,evaluation of information and found their preferences.Collaborative filtering recommendation is currently widely used recommendation algorithm,because of its relevance with the specific field is relatively weak,so the electricity providers,news,reviews,music and other industries have achieved relatively successful application.Simply put,you have a similar user preferences like to recommend to you.The main contents of this paper are as follows: The structured text processing of the user’s comment text,extracting the eigenvalues by the Chinese word segmentation,extracting the tendency word bag by the statistics of the word frequency,and personalized recommendation to the potential users.In this paper,we compare the user-based collaborative filtering recommendation algorithm and the user-based collaborative filtering recommendation.In the project-based personalized recommendation algorithm,the similarity calculation is optimized,and the similarity calculation method is proposed.The similarity of the scores is calculated based on the similarity of the emotional tendencies,the similarity based on the time series,and the recommended formulas are generated.Experimental results show that compared with the traditional similarity algorithm,the proposed algorithm improves the efficiency and improves the quality of the proposed algorithm. |
