Epidemic Disease Expert System

Abstract

The growth in the epidemic disease has created threat to human population. Many of the epidemic diseases are known to be sensitive to changes in the climate and short-term fluctuations in the weather. Recent outbreaks of Dengue, Chikungunya in India and Ebola in Africa have shown the importance of monitoring and understanding the public sentiments on disease outbreaks. These expert systems are fed with relevant knowledge and techniques to infer the result to make decisions. The paper focuses on extracting the web content like e-newspapers, Health related articles specific to the study of the disease occurrence and provide prior information of the outbreaks. The knowledge base is fed with the authenticated data collected from the field of medicine that provides information about the relationships between the diseases, symptoms, and medications. The system uses WebCrawler to crawl across the initial seed URL’s. The developed crawler supports the feature of multithreading crawling, content extraction and duplicate links detection. Around 3000 links are crawled by the WebCrawler. The dynamicity of the DOM structure of each e-newspaper is also taken care. Out of which few of the potential links related to the epidemic disease are taken for the study. Sentiment analysis of the curated newspaper article is demonstrated with real time data. The summary of the article, article published date, article title that are related to the epidemic disease is extracted. Pre-processing of data is done which includes parsing, tokenizing, lemmatizing and chinking of data. The Naïve Bayes Classifier is applied which results in the generation of score. The detailed view of the Disease related information is displayed to the user.