Research And Implementation Of APP Software User Behavior Analysis Method Based On User Comments

Faced with the rapid development of mobile Internet,APP software in the field of software rapidly rises as a new product.APP software has attracted much attention both in people's daily life and national security.Whether enterprises and individuals want to have a APP software to meet their own need.How to evaluate the quality of APP software by the user's point of view has become an urgent problem to be solved.Based on the user comment,the user behaviors of APP software are an indispensable part of APP software quality evaluation.However,due to the freedom and randomness of the network comments,which leads to the uneven quality of user comments,and the different needs of APP software users make it difficult to analysis the APP software user behavior,at last,this causes the difficulty with the evaluation of APP software quality.In order to evaluate APP software quality better,this paper is based on APP software user comments can obtain relevant information to study the users behavior analysis.We do five major as follows:(1)Define the user behaviors of APP software.Through the comprehensive analysis of APP software information,user's comment information and user's mark,we define that the user behaviors of APP software should include the relativity,consistency,patterns and credibility of user comment.(2)The relevance and consistency of user comment of APP software are judged.First,we judged the relativity of user comment by analyzing the relationships among user's comment information and APP software information.Then,through analyzing the grammar relationships among the feature-words,adverbs and the feature-sentiment-words in APP software's feature-sentiment-word-pairs,the user's emotional tendency about APP software is quantified combining with the dictionary and the network sentiment words,the consistency of APP software's user comment is judged by comparing this score and the user's mark.(3)This paper analyzes user comment patterns of APP software.Firstly,the user's comment information is classified based on the comment feature of APP soft-ware.Then the combination of part-of-speech for each type of user's comment information and the emotional tendency degree of user's comment information are analyzed,the user's comment patterns are analyzed.(4)The credibility of user comment for APP software was calculated.First,the credibility of user comment was modeled,and then the weights of different influence factors are analyzed and calculated in order to calculate the credibility of user comment.(5)In order to verify the correctness of the related theory and analysis method of user behavior,on this basis,the design of a prototype system is implemented and experiments are made to prove feasibility and effectiveness of this method.