Medical Data Mining using Soft Computing Technique

Abstract

Dengue fever, a human viral pathogen transmitted by mosquito, is a tropical infectious newlinedisease. Dengue diagnosis is not always possible in all medical centers, especially in newlinerural areas where assistance and care are reduced due to the lack of advanced dengue newlinediagnostic equipment. Early dengue disease signs and symptoms are also unspecific newlineand overlap with the other infectious diseases. In a small number of cases, dengue newlinedisease can be life-threatening and delay in diagnosis can increase the mortality risk. It newlineis therefore important to detect the dengue disease at very early stage. Hence the newlinedevelopment of diagnostic system for dengue is proven as a key research area in the newlinefield of biomedical informatics. newline Researchers have developed several methodologies to support the medical newlinediagnosis of dengue disease using artificial intelligence. Over the years, soft computing newlinehas played an important role in the diagnosis of such kind of disease and computer newlineaided systems ease the decision-making process for a doctor. With the advent of soft newlinecomputing technologies and the use of intelligent methods and algorithms provide a newlineviable alternative for vague, uncertain and complex diagnosis of dengue. newline In this work, soft computing and data mining technologies have been used to newlinedevelop models for early diagnosis of dengue disease. Various Machine learning newlineapproaches including the Artificial Neural Network (ANN), various Decision Trees newline(DT s) and Naive Bayes (NB), Particle Swarm Optimization-Artificial Neural Network newline(PSO-ANN), Support Vector Machine (SVM) and Genetic Algorithm (GA) have been newlineused. Also, dataset obtained from patients registered in various hospitals have been used newlinefor training and validating the aforesaid methods. This unique dataset is useful to newlineresearchers and practitioners working in dengue disease treatment and diagnosis. newline