Artificial Neural Network Based Categorization of Diabetes Mellitus and its Comparison with K-Nearest Neighbors Algorithm

The categorization of diabetes mellitus in humans with a Novel Artificial Neural Network (ANN) algorithm and it’sassessment with K-Nearest Neighbors to improve the performance in terms of precision and accuracyis proposed in this paper. Kaggle, a repository for machine learning algorithms has been used as a source of Pima Indian diabetes dataset. This dataset having 8 attributes in 6 different categories are considered as input attributes for the classifiers. 20 samples have been used for the testing of both the Novel artificial neural network algorithm and K-nearest neighborsAccuracy and precision of Artificial neural network algorithm is calculated and compared with K-nearest neighbors algorithm. The significance value is less than 0.05. The G power is taken as 0.8. In diabetes mellitus classification of the considereddataset, the proposed work shows that performance of the Novel artificial neural network algorithm is better as compared to the K-nearest neighbors algorithm.