Role of K-nearest neighbour in detection of Diabetes Mellitus

Abstract

Diabetes is one of the enduring and continuing illness in the world. According to world health organisation, it was approximately 104 million were suffering from diabetes in 1980 and in year 2014, the figure has risen to 422 million in the world and is expected to double by year 2030. In this paper, we have applied supervised K-nearest neighbour machine learning algorithm on PIMA Indians diabetes dataset. K- nearest neighbour algorithm works on the similarity between presented data and already stored data. We have shown that after the application of proposed algorithm, accuracy has risen from 70.1 % to 78.58% which is an increase of 8.48%.