Extracting Insights About Deep Learning Posts From StackOverflow

Abstract

In recent years, the field of deep learning has garnered more interest mainly due to the availability of vast amount of data and exponential improvement in processing power of computing devices. To share and expand their competency in a field like deep learning, Question and Answer sites provide essential support for developers and researchers. In this paper we investigate and extract data from a popular Q & A site, Stack Overflow to gather insights about the questions being asked and the corresponding responses on the topic of deep learning based on a set of tags. We gathered data related to posts by querying the Stack Exchange Data Explorer regarding information about the users and votes for each posts. We examined them to understand about how the designation and the reputation of user affect the quality of answers and how much time does it usually take to have an acceptable answer. Through our research, we attempted to answer 3 important research questions which we address in detail related to the relationship between the designation of the users and the reputation of the users on the quality of the question and answer posts. Furthermore, we derive insights regarding how quick one should expect good quality answers to arrive when a deep learning question has been posted.

See PDF for complete report.

Built With

• python