Project work submitted as part of the MSc Statistics course on Bayesian Machine Learning at the London School of Economics and Political Science
With the rise of online communities where labeling of web content got indispensable, automatic tag recommendation has arisen interest in recent machine learning literature. This paper aims to present and compare competing methods to predict tags of questions asked on Stack Overflow in a non-technical manner. We start by an introduction into the topic and related work, and proceed by explaining crucial text pre-processing steps. We then approach the problem of tag recommendation from two perspectives: (1) tag recommendation as classification problems and (2) as clustering problems.