The Design And Implementation Of The Public Scoring And Reviewing System Based On Deep Learning

The development of the information age change rapidly,with the further development of mobile Internet,big data era is coming,the internet is now filled with all kinds of data,people can easily obtain the required information,but the problem is that the mess of data makes it difficult to identify what is the real one.This situation is particularly prominent when it comes to commercial interests,the merchants tend to exaggerate or even fake their own product information to make customers to buy.Network public scoring and reviewing system is generated to solve this problem.The system takes advantage from this point by providing the vast number of scores and comments from consumers,to help users make more accurate judgments.Among them,the overall score is the most direct and most popular reference information.But the overall score will still be affected by some scores from customers with special purpose,and become distorted.This article focuses on solving this contradiction.With the use of depth learning algorithm from artificial intelligence field to optimize the scores,this article completed the entire structure of public scoring and reviewing system model,as weel as the implementation and verification.The result shows that the system can achieve a higher matching between the final presentation of the overall score and the selected standard score,thus brings the customers more objective and more accurate value of the product.The research work in this paper including:1.This paper analyzes the history and status of the research on the public scoring and reviewing system,studies the public scoring and reviewing system in current market,and found that the existing public scoring and reviewing systems are troubled with problems like reviews from users hired by merchant,extreme rating from users,lots of comments and ratings from newly registered and non-experienced users,other defects or deficiencies.This paper put forward through the application of artificial intelligence algorithm to extract the essential characteristics of user data from the mass comment data,make the level of weight correction according to the extracted features,thus optimizes the overall scores.By doing that,a new generation of solution for public scoring and reviewing system has come up.2.This paper studies the popular machine learning algorithm which can be applied to the data set,and compares and verifies the performance of different algorithms in the system.The data set is transformed into the form of image,which makes the data set in the convolution neural network algorithm to achieve an unexpected high accuracy.3.This paper analyzes the system requirements,completed the overall system architecture design.The system uses the B/S multi tier structure design,the front end is based on the JavaScript+HTML5+CSS technology realization,the back end is based on the Tensorflow+Python technology realization.The system is composed of items browsing,user scores and reviews,items management,system maintenance and other modules.System also provides an extension of the functional interface which can be compatible with a more common application scenarios,such as the combination of online payment and location-based services,online shopping,group purchase and other functions.4.This paper implements the general public scoring and reviewing system.It completed the design and implementation of user scores and reviews module,as well as the design and implementation of the optimization module based on user review data,and the front-end UI which can be adapted to different platform.The paper designs and implements the interface effect which can enhance the user experience,such as dynamic effect,carousel effect transitions.5.At last,this paper verifies the main function of the system.Verify the availability of the system to review the actual user scores.Through the comparison of the same data with different algorithms,the results show that the system is effective to improve the accuracy of the evaluation of the goods.The availability of cross platform solution is verified,and the applicability of the system is improved.This paper verifies the feasibility of optimizing the user scoring data through the research of the public scoring and reviewing system.The application of the system enhances the user experience and improves the accuracy of the review.