Article Per Year (5 Year)
p-Index From 2020 - 2025
0.23
P-Index
This Author published in this journals
All Journal Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer
Maulana Syahril Ramadhan Hardiono
Fakultas Ilmu Komputer, Universitas Brawijaya
Computer Science & IT Control & Systems Engineering Education Electrical & Electronics Engineering Engineering

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Rekomendasi Pengambilan Asuransi Kecelakaan bagi Driver menggunakan Improve K-Means dengan Inisialisasi Centroid berbasis Sum Square Error dan K-Nearest Neighbor Maulana Syahril Ramadhan Hardiono; Bayu Rahayudi; Dian Eka Ratnawati
Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer Vol 5 No 11 (2021): November 2021
Publisher : Fakultas Ilmu Komputer (FILKOM), Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Carelessness in driving can result in accidents that can harm yourself and others. To reduce losses caused by accidents, one can use insurance services. However, there are still many Indonesians who have not used insurance services due to high premiums. From these problems, a solution is needed to recommend someone to take insurance from the level of driving experience. One of the methods adopted from this research is to improve k-means with centroid initialization based on sum square error and k-nearest neighbor. The data used in this research are 600 data safe drivers, consisting of numeric data and categories. The recommendation process begins with clustering using k-means with the initialization of centroid sum square error which produces an average of 1000 iterations, which is 1660.64. From the SSE process, the optimum centroid was obtained and continued with the k-means process with k-prototype. The use of k-prototype is due to the data consisting of numeric and categorical data. After the clustering process, the distance between the tes data and the centroid is calculated to find the closest cluster. Furthermore, the classification process is carried out on the nearest cluster in order to make the classification process more efficient. The classification process uses KNN which aims to determine whether data is included in the recommendation category and not. From the tesing process the improve k-means method produces the highest average value of 72.83% accuracy, while the KNN method produces an average accuracy value of 47.5%. The results of these calculations use the value of k = 3 and k-fold on cross validation of 5.
Co-Authors Bayu Rahayudi Dian Eka Ratnawati