Prediction of Infectious Diseases based on Age and Gender using OCSVM-CDT

Abstract

Dengue, malaria, pneumonia and typhoid are the most notable infectious diseases because it causes terrible effects on people. These types of diseases affect particular age groups of people. Researchers face so many difficulties in finding the accurate affected age patterns of individual diseases. So, this study focused on age-wise and gender-wise occurrence of infectious diseases like dengue, malaria, pneumonia and typhoid. It is most useful to the doctors and health analyst to more concentrate on predicted age groups of people to reduce the death ratio. This prediction will create the disease alert system for the Madurai district. It will very helpful for the public. Machine learning is the most efficient technique to solve healthcare difficulties. The main aim of the study is to predict the age and gender distribution of these diseases using machine learning approaches. This study proposes a hybrid model One- Class Support Vector Machine – Chi Square Automatic Interaction Detection Decision Tree (OCSVM-CDT) for classifying the age and gender-wise occurrence of infectious diseases like dengue, malaria etc. Accuracy, precision, recall, and F1-score are the metrics used to evaluate the performance of this new hybrid technique. Five machine learning algorithms Such as Neural network, Decision tree, Support vector machine, K-means Chi-Square Automatic Interaction Detection Decision Tree (KM-CDT) and One-Class Support Vector Machine – Decision Tree (OCSVMDT) are compared with the proposed hybrid model. The proposed hybrid OCSVM-CDT performs better when compared to other machine learning models.