Project

Summarization, Opinion Mining and Pattern discovery from the health posts

Online health communities and blogs offer a large variety of useful medical information for users such as patients, medical practitioners and system administrators. In this work, we collect the real time data from the patients that are posted on the popular health related websites. This data may include information related to drugs like side-effects and the user’s opinion. The proposed system summarizes the user posts based on the specific date range which is helpful for the end users. Association rule mining is used to render the effective patterns on the triad ‘drugs-symptoms-medicines’. These patterns are advantageous for the knowledge discovery process. In addition, Opinion Mining (OM), also known as Sentiment Analysis (SA) which is a Classification process is used in this work to classify the users based on the ‘emotional state of mind’. The Classification process is performed based on the specific words. Also, symbols and special characters in the user posts are considered in this project which is mentioned as a future work in the proposed paper. Dynamic data (not static) is used for implementing this project. Hence, the system predicts the patterns for the dynamic data. Also, a plot of the number of satisfied patients versus number of depressed patients is provided based on a particular date range. 35 different types of graphs can be viewed in 2D as well as 3D. These operations are based on the data mining techniques and built on .Net framework.

Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.