A Study on Cuckoo Search based Hybrid Neural Network for Medical Data Classification

Abstract

Machine learning techniques are being increasingly used for detection and diagnosis of diseases for its accuracy and efficiency in pattern classification. In this paper, cuckoo search based hybrid neural network is proposed for classification of medical data. A tiny medical dataset called Dermatology, as well as three other benchmark datasets, are categorised and used for the study in this paper. Hybrid neuro fuzzy networks, such as cuckoo search based functional link neural fuzzy network (CSFLNFN) and cuckoo search based multilayered perceptron, were used in the study (CSMLP). These classifiers' performance is compared to that of Naïve, Bayes and KNN classifiers. To lower the dimension of these datasets, PCA is employed as a feature extraction approach, and the results obtained before and after feature extraction are compared.