Optimized Regression Neural Network For Classification Of Diabetes In Big Data Environment

Main Article Content

N.V.Poornima , Dr.B.Srinivasan , Dr.P.Prabhusundhar

Abstract

One of the main problems in the human body is diabetes. Diabetes is one of the critical health issues in the human body it causes human life severely. In women, the diabetes rate is more and higher as compared with the others. It severely affects pregnant ladies while they are affected by diabetes. This can be affected the daily routine. The one and only method are to take dialysis at the beginning stage. The famous doctors check the patients and provide the precautions and the medicines for that. They handle the patients affected by diabetes. To solve this problem several methods and techniques are used in the machine learning algorithm under the big data environment. This algorithm provides good prediction results by taking the data from the dataset. In this diabetes prediction, the machine learning algorithms combine with the big data environment. In the previous methods, the k-means clustering and cuckoo search is used in optimization. In this existing system, its accuracy and performance are high but it takes a huge time to computational process. To solve this problem optimization based on machine learning is applied. Here the dataset is downloaded from the Indian diabetes dataset. For removing the unwanted and missing data in the dataset we use pre-processing and clustering. And the feature reduction is done in the use of glow-worm optimization. The glow-worm is the fastest method compared to the cuckoo search optimization. The classification of diabetes is done with the help of feed-forward neural networks. The aim of the classification is that the user misunderstanding the values after removing the unwanted data. The whole process is realized in the MATLAB R 2018a environment and evaluated in terms of accuracy, precision, recall, F-measure, and Matthew correlation coefficient. This approach outperforms all other existing techniques with an F-measure of 98.6%.

Article Details

Section
Articles