A Supervised Machine Learning Study of Online Discussion Forums about Type-2 Diabetes

As an instance of online communities, online diabetes discussion forums mirror these characteristics and seem to track the growing impact of diabetes on individuals around the world. In this paper, we first systematically collected texts
from online discussion forums about diabetes and then applied supervised machine learning techniques to analyze the online conversations. In order to analyse these online textual conversations, we have chosen four domain specific models (Emotions, Sentiment, Personality Traits and Patient Journey). As part of text classification, we employed the ensemble learning method by using 5 different supervised machine learning algorithms to build a set of text classifiers by using the voting method to predict most probable label for a given textual conversation from the online discussion forums. Our findings show that there is a high amount of trust expressed by a subset of users and these users play a vital role in supporting other users of the online discussion forums about diabetes.