Health adviser: Social question and answer system using datamining

Abstract

Question and answer systems play a crucial role in providing answers to users that seek it, primarily in the case of non-factual questions. A Social Q&A system aims to integrate properties of social networks like similar interests of the users and mutual trust between within a set of friends which makes the reliability of certain answers more prudent. In simple terms, it aims to find a expert that can provide the right answers, ideally within the user’s social circle. Our project endeavors to serve pregnant mothers in answering their private or non-private questions regarding their pregnancies. Our system provides a platform for the to-be mothers to ask their questions to experts who can suggest suitable advises for their benefit. The user also has the added advantage of accessing information from experts who belong to their social circle thereby reducing the anonymity of the expert. We have compared the accuracy of few machine learning algorithms that includes Decision Tree (DT), Support Vector Machine (SVM), K-Nearest Neighbor kNN), Logistic Regression (LR), ZeroR (the baseline classifier). Our model has an accuracy of over 75%, while demonstrating robustness across learning algorithms.