Presentation Title

Using Machine Learning to Predict Response Time on Stack Overflow

Format of Presentation

Poster to be presented Friday March 31, 2017

Abstract

This project uses machine leaning algorithms to predict the response time to questions posted on the Stack Overflow website. Stack Overflow is the largest online community for developers and software programmers to learn, share their knowledge, and post technical questions. We created a training dataset of questions posted over a three-month time period. Then, we used feature extraction/selection techniques to identify the most important features that could be used to create unique profiles for common questions. We were able to find common patterns for similar questions posted during the same time period. We conducted several experiments to develop predictive models that can learn from historical data in the training dataset and predict the response time of future questions. We were able to improve the classification accuracy from 39% to 89% using supervised machine learning approach. We used cross-validation methods to validate the accuracy of our machine learning algorithm.

Department

Computing Science

Faculty Advisor

Haytham El Miligi

This document is currently not available here.

Share

COinS
 

Using Machine Learning to Predict Response Time on Stack Overflow

This project uses machine leaning algorithms to predict the response time to questions posted on the Stack Overflow website. Stack Overflow is the largest online community for developers and software programmers to learn, share their knowledge, and post technical questions. We created a training dataset of questions posted over a three-month time period. Then, we used feature extraction/selection techniques to identify the most important features that could be used to create unique profiles for common questions. We were able to find common patterns for similar questions posted during the same time period. We conducted several experiments to develop predictive models that can learn from historical data in the training dataset and predict the response time of future questions. We were able to improve the classification accuracy from 39% to 89% using supervised machine learning approach. We used cross-validation methods to validate the accuracy of our machine learning algorithm.