A genetic algorithm, information gain and artificial neural network based approach for hypertension diagnosis

Chronic disease are the leading cause of increase in mortality rate globally, In this study we cosidered hypertension as case study. Diagnosis and screening of hypertension plays a significant role in the prevention and reduction of the onset of cardiovascular diseases as well as the improvement of quality of life. It is of extraordinary worth to make sense of hypertension-related risk factors and further further set up a model for the forecast of hypertension with the identified risk factors. Proposed approach introduces a model that seeks for the improved performance of the classifier in terms of accuracy of classification model, number of features to build the model etc. This paper focuses on a comparative study of two appraoches to achieve improved performance. Information gain based attribute subset selection approach is compared with Genetic algorithm based appraoch to discover optimal subset of attributes that adds more to the classification. Then these attributes are given as input to multilayer perceptron neural network with backpropagation algorithm. Experimental results shows that accuarcy achieved in prediction of hypertension by using information gain approach is 97.58% with 9 features where as Genetic algorithms gives prediction accuracy of 99.19%with 9 different features. Results indicates performance of classification model of medical diagnosis can be improved with proposed approach.