Penerapan Particle Swarm Optimization Untuk Meningkatkan Kinerja Algoritma K-Nearest Neighbor Dalam Klasifikasi Penyakit
Diabetes
Total View This Week1
Total View This Week1
Institusion
STMIK Bumi Gora
Author
Sutrisno, Sutrisno
Subject
T Technology (General)
Datestamp
2021-08-20 02:24:16
Abstract :
Diabetes is a serious problem that can cause complications, leading to death. Word Health Organization (WHO) diabetes killed 1.5 million people in 2012. From these problems many studies in the field of data mining that apply classification methods use to diagnose or identify disease diabetes. This study aims to determine the application and level of accuracy obtained by the KNN algorithm in the classification of diabetes and knowing the level of accuracy of the KNN algorithm after being applied by BPSO as a feature selection. This research was carried out in stages, namely dataset collection pima indian diabetes as a test dataset, pre-processing the data using min-max normalization and BPSO feature selection, dataset sharing using 10-fold cross validation method, classification using the KNN algorithm up to the testing process to determine the level of accuracy obtained. The results showed that the classification using the KNN . algorithm by implementing feature selection using BPSO to obtain the best accuracy that is equal to 77.214% with the obtained features that affect the pima . dataset namely Glucose, Blood Pressure, Skin Thickness, Insulin, BMI, Diabetes Pedigree Function, and Age while the classification uses the KNN algorithm without BPSO feature selection obtained a lower accuracy of 75%. With Thus the KNN algorithm with BPSO feature selection can be used to identify diabetes because it has a high level of accuracy better than the KNN algorithm alone. Keywords: Diabetes Classification, Particle Swarm Optimization, KNearest Neighbor
Diabetes is a serious problem that can cause complications, leading to death. Word Health Organization (WHO) diabetes killed 1.5 million people in 2012. From these problems many studies in the field of data mining that apply classification methods use to diagnose or identify disease diabetes. This study aims to determine the application and level of accuracy obtained by the KNN algorithm in the classification of diabetes and knowing the level of accuracy of the KNN algorithm after being applied by BPSO as a feature selection. This research was carried out in stages, namely dataset collection pima indian diabetes as a test dataset, pre-processing the data using min-max normalization and BPSO feature selection, dataset sharing using 10-fold cross validation method, classification using the KNN algorithm up to the testing process to determine the level of accuracy obtained. The results showed that the classification using the KNN . algorithm by implementing feature selection using BPSO to obtain the best accuracy that is equal to 77.214% with the obtained features that affect the pima . dataset namely Glucose, Blood Pressure, Skin Thickness, Insulin, BMI, Diabetes Pedigree Function, and Age while the classification uses the KNN algorithm without BPSO feature selection obtained a lower accuracy of 75%. With Thus the KNN algorithm with BPSO feature selection can be used to identify diabetes because it has a high level of accuracy better than the KNN algorithm alone. Keywords: Diabetes Classification, Particle Swarm Optimization, KNearest Neighbor
Download
book
BibTex
Latex, Jabref
cloud_download
Original Resource
url resource Institution
