A proposed approach to determining expertise level of StackOverflow programmers based on mining of user comments

Abstract

Abstract The popularity of Question–Answer websites such as StackOverflow, Ask and Yahoo! Answers is gradually increasing. Considering this increased popularity, the quality of the questions and answers is necessary to be taken into account. The reason for this necessity is that there could be many answers to any question, and some of these answers have a low level in terms of quality. The credibility and expertise of the questioner and the respondents in the field of the question is one of the solutions to get around this problem. In other words, individuals with a high level of expertise ask more difficult and high-quality questions in their field of expertise, and individuals with a high level of expertise can answer these questions appropriately. The present paper aims at finding the expertise level of the individuals based on available statistical data about questions and answers. For this purpose, two methods were tested. In the first method, which is performed through scoring in this study, the emphasis is on the scores of the questions and answers. The basic assumption of the scoring method is that the scores of the questions and the answers are related to each other, because individuals with a higher level of expertise raise questions with higher scores and, on the other hand, respondents also provide answers with higher scores. Therefore, it is possible to determine the level of expertise of individuals by examining the scores of the questions and answers. In the second method, which is named comment-mining here, the expertise of an individual is determined through scoring positive and negative words in the comments and ultimately obtaining a final score for the questions and answers of a user. The actual data on the StackOverflow website were used to perform these methods. The results of the scoring method show that there is no significant relationship between the scores of the questions and answers, so this method cannot be used to determine the level of expertise of users. But the results from the comment-mining method suggest that the method determines the expertise level of user to an acceptable level.