Classification of Infection Type Based on Leukocytes Examination Results Using K-Nearest Neighbor

Abstract

Leukocytes are blood cells that contain nuclei, also called white blood cells. Leukocytes have a role in the cellular and humoral defiance of organisms against foreign substances. Laboratory tests of blood samples greatly influence the diagnosis of a disease. Manual blood tests do have a low price but still have some weaknesses such as the length of time needed will be longer, because health practitioners must examine them carefully to avoid misinformation. To help overcome these weaknesses, a classification of types of infections was carried out based on the results of leukocyte examination. Classification is a grouping of data where the data used has a label or target class. So that the algorithms for solving classification problems are categorized into supervised learning. The purpose of supervised learning is that label data or targets play a role as a 'supervisor' or 'teacher' who oversees the learning process in achieving a certain level of accuracy or precision. The algorithm used in this study is K-Nearest Neighbour. The data used in this study as many as 2,098 results of complete blood tests taken from one hospital in Medan. This study resulted in a classification accuracy of 92%.