Authors : Jagadeesh Gobal and Subhashini Narayan
Abstract: Heart disease is the primary cause of death in all over the world and one of the primary diseases in developing countries. The disease identification and investigation needs a lot of prognosis and diagnosis with the high cost. To eliminate these burdens, we need to predict this disease in early stages with the help of machine learning techniques. The UCI repository heart data such as Cleveland, Hungarian dataset and Switzerland dataset for disease prediction. The proposed method consists of three stages: initially inconsistent, irrelevant and noisy data from the dataset using the fundamental fuzzy min. max. preprocessing, after that orthogonal locality preservation projection technique is applied to the processed data to reduce the dimension. At the second stage, chaotic cuckoo search algorithmetic approach is used for fitness evaluation, the combinations of cuckoo search algorithm, fuzzy and decision tree classifier can create a hybrid class. Information entropy algorithm will be sufficiently combined with cuckoo search algorithm achieves better classification accuracy. The result of classification accuracy of other existing algorithms is respectively 0.89, 0.93, 0.91. Consequently, the obtained results have shown very promising outcomes for the diagnosis of heart disease.
Jagadeesh Gobal and Subhashini Narayan, 2017. Hybrid Decision Tree Fuzzy Rule Based Classifier for Heart Disease Prediction Using Chaotic Cuckoo Search Algorithm. Journal of Engineering and Applied Sciences, 12: 7358-7366.