Surveying the impact of environmental conditions on diabetes using ensemble classifier based on genetic algorithm

Abstract

Medical data mining is considered as a new solution to analyze medical data and discover knowledge . Medical data mining has a high potential for discovering hidden patterns in medical data. In the present era, some studies have been conducted on the relationship between environmental quality and diseases , which have clearly indicated the impact of environmental quality indicators, such as environmental pollutants, on diseases. In this study, environmental conditions in diabetes were investigated based on medical data mining technique. Diabetes is considered as a global threat affecting human health. An ensemble classifier based on genetic algorithm(ECGA) method was designed to study the environmental conditions in diabetes. In the designed ensemble classifier, the decision tree, random forest, k-nearest neighbor, and naive bayes were used. It was found that ECGA was more accurate than the base classifier algorithms. In addition, three datasets were collected from different regions of Iran with different climatic conditions. It was found that environmental conditions can affect diabetes disease.