MACHINE LEARNING TECHNIQUE FOR EARLY PREDICTION OF HYPERTENSION

Abstract

Machine Learning (ML) has proved to be an invaluable tool in medical research today. It is a branch of artificial intelligence that has the ability to learn from complex data, identify patterns and make decisions with minimal or without human intervention. Efforts in this work is focused at developing an intelligent model for early prediction of hypertension because of the complexity associated with the pattern identification of the syndrome. The study acquired data pertaining to the syndrome from selected healthcare centres in the South West region of Nigeria, one of the largest city in Africa. A retrospective learning using ML analyses was conducted on the data. Three different intelligent classification algorithms namely; Gradient Boosting, Random Forest and SVM were used in the modeling. The datasets were trained and tested, and the performance of the algorithms were compared using standard performance metrics of accuracy and throughput. The results from the experiment indicates that the gradient boosting classifier outclassed other algorithms considered in the work with accuracy of 90.17%. The model has proved to be very suitable for early prediction of the syndrome and, will in no small measure assist the medical practitioners in providing quick and more effective way of detecting early the syndrome in patients. This will subsequently lead to timely treatment of the patients, thereby reducing its death rate in the society. Future work is focused at considering complex patterns of the syndrome and larger database.