Predicting Diabetes Mellitus Using Modified Support Vector Machine with Cloud Security

Diabetes mellitus is one of the major concerned diseases that cause a large number of deaths every year. It is considered as the chronic disease which is caused by an increase in blood sugar. If diabetes remains unidentified and untreated, it creates more complexities. So, the early prediction of diabetes can reduce the fatal rate of a human. The data mining concept assists to diagnose diabetes. Various research studies are presented with various data mining algorithms for early prediction and disease diagnosis but still with lack of accuracy. At the same time, mining the diabetes data in a secure manner is one of the critical issues. To recover this issue, this paper designs the new model for early prediction of diabetes with high accuracy. This research explores the enhanced principal component analysis for efficient feature extraction from the dataset. To achieve the highest accuracy of classification, it has proposed the machine learning algorithm, namely, modified support vector machine (MSVM) which is used to detect the diabetes disease at an early stage. The main contribution of this research is to mining the patient’s disease results in cloud security. For this security purpose, honey bee encryption and decryption algorithm is used. The performance measures of the proposed method are evaluated on various measures of accuracy, sensitivity, specificity, precision, and negative predictive value. Results obtained show the proposed MSVM classifier outperforms with the highest accuracy of 97.13%. We have compared the proposed methods with existing methods for proving our method has better performance.