Identifikasi Penyakit Tiroid Menggunakan Algoritma Modified K-Nearest Neighbor

Abstract

Thyroid disease is a type of cancer in which tissue cancer cells grow uncontrollably in the thyroid tissue. Diagnosing this disease is quite difficult because it has various symptoms due to the influence on the rise and fall of the thyroid hormone and increase the use of oxygen by the body's cells. Diagnosing this disease requires a thyroid examination by a doctor as well as the appropriate interpretation of clinical data. A doctor in diagnosing is limited by the age factor and time constraints causing a lack of interpretation of a patient's clinical data. In this study, thyroid disease was identified using a thyroid disease dataset sourced from the UCI Machine Learning Repository with the Modified K-NN algorithm. The Thyroid dataset is in Microsoft Excel Comma Separated Values File (.csv) format which consists of 21 attributes and 1 class with 2754 records. This study was conducted to determine the accuracy of the identification results using the Modified K-NN (MK-NN) algorithm. The result of system testing is that the system can identify thyroid disease datasets by using training datasets of 2754 and test data of 10, 50, 100, 200 and 300 records. The best identification accuracy value is 98,5 % on test data of 200 records.