Feasibility analysis for popularity prediction of stack exchange posts based on its initial content
Devaraj Phukan, Aayush Kumar Singha · International Conference on Computing for Sustainable Global Development · 2016
In technology related question and answer websites, people working on corresponding technologies post questions about the problems that they encounter and seek answers to them. The popular posts having greater views that stay active for longer periods of time are the ones that occur commonly for many users. The objective of this paper is to predict such posts at its arrival time. Using the prediction results, the problems that will gain popularity over time can be addressed before they become very common. The posts in question and answer websites have a longer lifespan than other online content such as news articles which are short lived and have a high user activity at the beginning that decrease with time. This indicates that the user activity in question and answer websites is not an accurate measure of popularity in an early stage. In this paper, we have used the initial content in a post as the predictors of popularity rather than using conventional methods such as shares, comments etc. We used three models to predict the popularity using the dataset of android.stackexchange.com, a popular question and answer website for android developers.