Hands-On Tutorial on Stack Overflow Question Tagging
This article was published as a part of the Data Science Blogathon.
Background
I won’t be lying if I assert that every developer/engineer/student has used the website Stack Overflow more than once in their journey. Widely considered as one of the largest and more trusted websites for developers to learn and share their knowledge, the website presently hosts in excess of 10,000,000 questions. In this post, we try to predict the question tags based on the question text asked on Stack Overflow. The most common question tags on Stack Overflow include Java, JavaScript, C#, PHP, Android amongst others.
Proper prediction of the tags is important to ensure that the questions are suggested to users having substantial experience in answers related to the topic suggested.
Data
The data was collected from the Keyword Extraction competition by Facebook on Kaggle (https://www.kaggle.com/c/facebook-recruiting-iii-keyword-extraction/data).
The training data consists of 4 columns- Id, Title,