An improved machine learning model for diabetes predication using DR image

  • Deepshikha Gupta, Aman Jatain, Sarika Chaudhary

Abstract

Diabetic retinopathy (DR), referred to as Diabetic. This retinopathy disease is continuously increasing in our country. Diabetic Retinopathy (DR) is an affects the blood vessels in the retina. It is due to the retina not receiving enough oxygen. Then after that our eyesight starts decreasing, and then later becomes the leading cause of blindness.  Many such rare a lot of research to reduce the disease is taking place in the biomedical field. In this way, many machine learning tools are being used in the field of biomedical for different patterns and diabetes retinopathy prediction. In addition, the current method of detecting diabetes retinopathy is a extremely laborious and time intensive task, which depends on the skills of the physician. Retinopathy is necessary to find out whether any human is diabetic or not and to deal with these problems. It is also very important to detect and diagnose diabetic retinopathy in primary stage, which can prevent blindness with proper treatment. In this paper, we have developed a model system that detects diabetic retinopathy from retinal images. Two machine learning algorithms have been used in this model system. And then provides results on comparing the accuracy of both. We tested our system with the largest dataset available on Kaggle Repository and then we will build a machine learning model. We have used two different types of classifier such as SVM and Logistic Regression and then we have got the accuracy value of SVM classifier is 0.93 and Logistic Regression classifier accuracy value is 0.94.  So we have analyzed that Logistic Regression classifier is much better as compare to SVM classifier.

Published
2020-06-01
How to Cite
Deepshikha Gupta, Aman Jatain, Sarika Chaudhary. (2020). An improved machine learning model for diabetes predication using DR image. International Journal of Advanced Science and Technology, 29(7), 8378-8387. Retrieved from http://sersc.org/journals/index.php/IJAST/article/view/24871
Section
Articles