Font Size: a A A

An Analysis Of Java Related Posts On Stack Overflow

Posted on:2020-05-22Degree:MasterType:Thesis
Country:ChinaCandidate:P ZhangFull Text:PDF
GTID:2428330590996804Subject:Software engineering
Abstract/Summary:PDF Full Text Request
Stack Overflow(often referred to as "SO")is a famous foreign programming platform of technical question and answer,on the platform,software developers ask questions about specific programming languages or development techniques,and other users use their domain knowledge and experience to provide answers.According to statistics,the percentage of questions and answers on Stack Overflow about Java development frameworks and the third-party development packages are pretty high.As one of the most popular programming languages today,Java provides a wealth of class libraries and API documentation.It will help software developers make better use of the language by analyzing Java related posts on Stack Overflow.Hence,in this article,we present a set of complete topic model analysis method,which is used for statistical analysis and topic model analysis of java-related posts on Stack Overflow.Our analytical framework is divided into four steps.(1)Data set retrieval: We first download the dataset from Stack Exchange Data Dump.Then we extract the data about Java.At last,we preprocess the dataset.(2)Java technology trend analysis: based on the number of posts on Stack Overflow,we carry on the statistical analysis to the current popular Java technology.We not only analyze some Java technology that popular in recent years though the trend of the posts number over time in order to understand the usage of these technology in recent years,but also compare some of the same type technology though the posts number in recent years,in order to understand the developers' attention of the same type technology in recent years.(3)Topic model analysis: First,we use the LDA(Latent Dirichlet Allocation)topic model to find the topic discussed by developers automatically.Then,we explore the relationship between topics and posts and calculate the relevant indicators of the topic according to the relevant attributes of the posts.(4)Analysis of experimental results: based on the analysis result of the topic model,we rank the probability matrix obtained by the topic model and answer the questions based on the experimental results,and draw valuable conclusions that the developers in this field are interested in.Combined with the above frameworks,this article helps developers understand the changing trends of different Java technologies and their usage in recent years,and helps developers understand the common problems and solutions of each Java technology,and then drawing valuable conclusions and topics of interest to developers in this field.These findings provide valuable experience for developers to use the Java language and technology in the future.
Keywords/Search Tags:Stack Overflow, Java, Topic Model, LDA
PDF Full Text Request
Related items