Implementasi Algoritma K-Medoids untuk Pemetaan Penyebaran Covid-19 di Surabaya

Dewi, Vivi Mentari (2021) Implementasi Algoritma K-Medoids untuk Pemetaan Penyebaran Covid-19 di Surabaya. Diploma thesis, Institut Teknologi Sepuluh Nopember.

Text Vivi Mentari Dewi_10611710000009.pdf - Accepted Version Restricted to Repository staff only until 1 October 2023. Download (1MB) | Request a copy

Abstract

Surabaya tercatat sebagai kota dengan kasus aktif Covid-19 tertinggi di Jawa Timur sebanyak 225 kasus berdasarkan data yang dirilis Provinsi Jawa Timur pada 01 Maret 2021. Salah satu langkah untuk meminimalisir kenaikan kasus yaitu dengan mengelompokkan wilayah berdasarkan jumlah kasus yang ada. Pemetaan yang ada sebelumnya hanya menampilkan data status pasien terkonfirmasi pada setiap wilayah yang diperbarui setiap harinya. Sehingga pada penelitian ini dilakukan pemetaan kelurahan di Surabaya yang akan dimasukkan dalam klaster berdasarkan kasus konfirmasi dalam perawatan, konfirmasi sembuh, dan konfirmasi meninggal menggunakan Algoritma K-Medoids. K-Medoids merupakan algoritma klasterisasi (unsupervised learning) dalam machine learning, pengembangan dari K-Means yang sensitif terhadap outlier. K-Medoids memiliki performa klasterisasi yang lebih baik untuk dataset dalam jumlah besar. Hasil klasterisasi menggunakan K-Medoids dengan nilai evaluasi Davies-Bouldin Index (DBI) sebesar 0,5666 diperoleh klaster optimum sebanyak 4 klaster. Klaster 1 (kasus konfirmasi meninggal tinggi) 35 kelurahan, klaster 2 dan 3 (kasus konfirmasi dalam perawatan, kasus konfirmasi sembuh, dan kasus konfirmasi meninggal rendah) 56 dan 49 kelurahan, dan klaster 4 (kasus konfirmasi dalam perawatan dan kasus konfirmasi sembuh tinggi) 14 kelurahan.
=========================================================
===================================
Surabaya was recorded as the city with the highest active Covid-19 in East Java with 225 cases based on data released by the East Java Province on March 1st, 2021. One step to minimizing the increase of cases is by grouping regions based on the number of existing cases. The previous mapping only displayed data on the status of confirmed patients in each region which was updated daily. So in this research, a mapping of urban villages in Surabaya was carried out which would be included in the cluster based on confirmed cases under treatment, confirmed cases recovered, and confirmed cases death using K-Medoids algorithm. K-Medoids is a clustering algorithm (unsupervised learning) in machine learning, the development of K-Means which is sensitive to outliers. K-Medoids has better clustering performance for large datasets. The results of the analysis showed that the urban villages with the highest number of deaths were Karah and Kutisari with 3 cases. The results of clustering using K-Medoids with an evaluation value of the Davies-Bouldin Index (DBI) of 0.5666 obtained the optimum cluster of 4 clusters. The first cluster (high confirmed cases of death) 35 urban villages, second and third clusters (low confirmed cases of under treatment, recovered, and death) 56 and 49 urban villages, and fourth cluster (high confirmed cases of under treatment and recovered) 14 urban villages.

Item Type:	Thesis (Diploma)
Uncontrolled Keywords:	Covid-19, Davies-Bouldin Index (DBI), K-Medoids, Machine Learning, Unsupervised Learning
Subjects:	H Social Sciences > HA Statistics > HA29 Theory and method of social science statistics
Divisions:	Faculty of Vocational > 49501-Business Statistics
Depositing User:	Dewi Vivi Mentari
Date Deposited:	14 Aug 2021 01:51
Last Modified:	14 Aug 2021 01:51
URI:	http://repository.its.ac.id/id/eprint/86298

Actions (login required)

View Item