Abstract
This paper proposes a new training algorithm for artificial neural networks based on an enhanced version of the grey wolf optimizer (GWO) algorithm. The proposed model is used for classifying the patients of diabetes disease. The results showed that the proposed training algorithm enhanced the performance of ANNs with a better classification accuracy as compared to the other state of art training algorithms for the classification of diabetes on publicly available “Pima Indian Diabetes (PID) dataset”. Several experiments have been executed on this dataset with variation in size of the population, techniques to handle missing data, and their impact on classification accuracy has been discussed. Finally, the results are compared with other nature-inspired algorithms trained ANN. EGWO attained better results in terms of classification accuracy than the other algorithms. The convergence curve proved that EGWO had balanced the local and global search abilities because it was faster to reach better positions than the original GWO.
Highlights
Diabetes is a common health problem and is often described by professionals as diabetes mellitus (DM)
This paper proposes a new training algorithm for artificial neural networks based on an enhanced version of the grey wolf optimizer (GWO) algorithm
The results showed that the proposed training algorithm enhanced the performance of advanced neural networks (ANN) with a better classification accuracy as compared to the other state of art training algorithms for the classification of diabetes on publicly available Pima Indian Diabetes (PID) dataset
Summary
Diabetes is a common health problem and is often described by professionals as diabetes mellitus (DM). The US Centers for Disease Control and Prevention (CDC) predicted a 23% increase in type-II DM cases for nine years straight (2001–2009). The availability of data and the rise of computational capabilities have encouraged scientists to analyze clinical data to find answers to these pressing problems This is done using data-mining techniques to discover useful information from large health datasets. Aljumah et al (2013) concentrated on predictive analyses of diabetic treatment using regression-based data-mining techniques. In a study by Bashir et al (2016), a new hybridization of different classifiers was proposed for the prediction and classification of diabetes This hybridization approach was proposed to overcome the issues associated with each of the classifiers. We introduce the methodology, results, and discussion, and we offer some closing thoughts in the conclusion
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
