Research On The Impact Of User Negative Reviews On App Software Update

Information technology and artificial intelligence is imperceptibly changing people's lives.Smart phones have become indispensable in learning and working.App software is becoming more and more popular.After downloading and installing the software,the user will comment on his reviews in the app store,and the App software developer will update the software periodically and publish the update logs.User's negative reviews contain a large number of user needs and opinions and suggestions on App software.It is important to study how software developers can use negative reviews information to update App software.In this paper,by crawling the negative reviews and update log information of App software,extracting the features of these information,and defining the update mode and update node according to the strength of reviews and log features.By calculating the update effectiveness and timeliness of all update nodes,we cluster five update modes,and make a series of analysis on update mode.The specific research in this paper includes the following contents:(1)Analyzing the current status of App software in major application stores,writing web crawler program,crawling the negative reviews and update log information of App software in the top 100 free rankings of Apple App Store,and statistical analysis of the original data,mainly visualizing the number of different star reviews and updates.(2)Text pre-processing is carried out on the original data,including spam review filtering,Chinese word segmentation and text stop word filtering,and feature extraction is performed on the pre-processed data using topic model.This paper defines the update mode and update node,describes the comment feature and update feature intensity,gives the calculation method of update effectiveness and timeliness,and calculates the response of developers to the feedback feature in user's negative reviews by using effectiveness and timeliness.Five update modes are mined by using K-means clustering algorithm,and these five update patterns are described.(3)According to the five update modes obtained from mining,the relationship between the update mode of App and rankings is analyzed by multiple linear regression.It is found that the lower the effectiveness of update,the lower the ranking of App software,the higher the effectiveness of update,and the higher the ranking of App software.The low timeliness of update is not conducive to the improvement of App software ranking.By analyzing the update mode and negative reviews on the relationship between one-star,two-star,three-star,and total number found that there is no correlation between the update mode and the number of negative reviews.By analyzing the similarity of update modes,it is found that the update modes of certain features of App mainly depend on software developers.By analyzing the stability of update modes,it is found that software developers tend to choose different update modes when responding to features.The conclusion of this paper can help developers understand the rules of responding to negative reviews,and has certain significance for developers to improve the quality of App software.