Diabetes Type2 Patient Detection Using LASSO Based CFFNN Machine Learning Approach
Sandeep Tiwari, Neetesh Gupta, Pranay Yadav
Abstract
Type 2 diabetes is a long-lived condition that prevents the insulin human system properly using. Insulin sensitivity is a condition that affects patients with type 2 diabetes. This type of diabetes seems to be more common among people in their early and middle years of life. In the proposed hybrid model is the combination of cascaded feed forward neural network (CFFNN) and Lasso Regression method. Lasso regression has always been a kind of regression in deep learning, which participate in training with the decision of functions. The absolute magnitude of the regression coefficient is prohibited. For the simulation of proposed method utilize MATlab (r2018b). For the analysis Diabetes Type2 Patient use Pima Indian and UCI data sets. The proposed hybrid approach shows better outcomes as compare to other recently presented methods in terms of accuracy and other performance parameters.