Prevalence of hypertension: predictive analytics review

Abstract

Hypertension is one of the non-communicable disease (NCD) that is classify as a global health risk with many critical health cases. Malaysia raise the same concern of the increasing NCD health problem. This paper aims to study the techniques used in predictive analytics namely healthcare and identify the factors of prevalence on hypertension. This review would give a better understanding of proper techniques and suggest the technique commonly used in predictive analytics especially for medical data and at the same time provide significant factors of prevalence hypertension. A total of 27 papers reviewed, several techniques on predictive analytics in healthcare are neural network, decision tree, naïve bayes, regression and support vector machine. The rise of economic growth and correlated socio-demographic have cause rise in hypertension problem over past years. The factors of hypertension depicted in this review namely gender, age, locality, family history, physically inactive and unhealthy life style not conform to any boundaries thus far. Thus, the choice on the technique and hypertension factors for predictive analytics is significant to come out with the significant predictive model. The predictive model on prevalence of hypertension may predict the severity of adult having hypertension in future work.

Simulated trial and error experiments on productivity

“…As most ML models are not restricted by linear specifications [9], accurateness is on the fly with the cost of harder interpretability and computation power [10]. With growing applications to predict in construction labor productivity [11], real estate [12], medical [13], epidermiology [14], finance [15], road maintenance [16], and other more. Random forest, gradient boosting, decision tree and neural network based models were proved to be outstanding choice for accuracy.…”