Abstract
Diabetes is more severe inwomen, according to various medical reports and surveys. Sometimes diabetes is difficult to identify due to various common symptoms, such as headache, fatigue, slow healing of cuts and blurry vision. Thus, this paper introduces novel big data and classification techniques such as effective map reducing technologies are used to recognize the diabetes. Initially, the data were collected from a large dataset, and the map reducing concept is applied to compose the small chunk of data efficiently. Following this process, the noise present in the collected dataset is removed using the normalization process. After that, the statistical features are selected using the ant bee colony approach that uses the ant characteristics such as wandering. The selected features are trained with the help of the support vector machine with multilayer neural network. The trained or learned features are efficiently classified using the associated neural network, and the efficiency of the system is evaluated with the help of experimental results in terms of error rate, sensitivity, specificity and accuracy.