Estiningtyas, Raissa Undita (2026) Perbandingan K-Means, Fuzzy C-Means, Fuzzy Rough C-Means, Dan Firefly–Fuzzy Rough C-Means Untuk Klasterisasi Data Sampah Tingkat Provinsi Di Indonesia. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5002221108-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5002221108-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (1MB) | Request a copy |
Abstract
Permasalahan pengelolaan sampah di Indonesia menunjukkan kompleksitas yang tinggi akibat perbedaan karakteristik wilayah, jenis sampah, dan sumber sampah tingkat provinsi. Untuk memahami pola tersebut, diperlukan metode analisis data yang mampu menangani ketidakpastian dan batas klaster yang tidak tegas. Penelitian ini bertujuan membandingkan performa metode K-Means, Fuzzy C-Means (FCM), Fuzzy Rough C-Means (FRCM), dan Firefly-Fuzzy Rough C-Means (FA-FRCM) pada analisis data sampah tingkat provinsi di Indonesia. Data penelitian diperoleh dari Sistem Informasi Pengelolaan Sampah Nasional (SIPSN) periode 2019-2025 yang meliputi data timbulan sampah, jenis sampah, dan sumber sampah pada tingkat kabupaten/kota yang diagregasi menjadi tingkat provinsi. Tahapan penelitian meliputi pra-pemrosesan data, penanganan missing value, agregasi data, normalisasi, tuning parameter, penentuan jumlah klaster optimal menggunakan elbow method, proses klasterisasi, evaluasi hasil, uji statistik, serta visualisasi klaster. Evaluasi kualitas klaster dilakukan menggunakan Silhouette Index, Dunn Index, Symmetric Purity, dan Xie-Beni Index. Perbedaan performa antarmetode selanjutnya dianalisis menggunakan Friedman Test dan Nemenyi Test. Hasil penelitian menunjukkan bahwa FRCM memperoleh peringkat akhir terbaik berdasarkan kombinasi Silhouette Index, Dunn Index, dan Symmetric Purity, dengan rata-rata peringkat sebesar 1,8683. Sementara itu, FCM memperoleh performa terbaik berdasarkan Xie-Beni Index dengan rata-rata peringkat sebesar 1,0200. Hasil Friedman Test pada seluruh metrik menghasilkan p-value kurang dari 0,05, yang menunjukkan adanya perbedaan performa yang signifikan antarmetode klasterisasi. Visualisasi menggunakan t-SNE dan peta intuitif menunjukkan bahwa karakteristik sampah tingkat provinsi di Indonesia tidak bersifat homogen dan beberapa provinsi mengalami perubahan klaster selama periode pengamatan. Secara keseluruhan, FRCM menjadi metode paling kompetitif dalam menghasilkan struktur klaster berdasarkan kombinasi ketiga metrik evaluasi utama.
====================================================================================================================================
Waste management in Indonesia is highly complex due to differences in regional characteristics, waste types, and waste sources at the provincial level. To understand these patterns, a data analysis method capable of handling uncertainty and indistinct cluster boundaries is required. This study aims to compare the performance of K-Means, Fuzzy C-Means (FCM), Fuzzy Rough C-Means (FRCM), and Firefly-Fuzzy Rough C-Means (FA-FRCM) in analyzing province-level waste data in Indonesia. The data were obtained from the National Waste Management Information System (SIPSN) for the 2019-2025 period and included waste generation, waste composition, and waste source data at the regency/city level, which were aggregated to the provincial level. The research stages consisted of data preprocessing, missing value handling, data aggregation, normalization, parameter tuning, determination of the optimal number of clusters using the elbow method, clustering, performance evaluation, statistical testing, and cluster visualization. Cluster quality was evaluated using the Silhouette Index, Dunn Index, Symmetric Purity, and Xie-Beni Index. Differences in performance among the methods were further analyzed using the Friedman Test and Nemenyi Test. The results showed that FRCM achieved the best final ranking based on the combined evaluation of the Silhouette Index, Dunn Index, and Symmetric Purity, with an average rank of 1.8683. Meanwhile, FCM achieved the best performance based on the Xie-Beni Index, with an average rank of 1.0200. The Friedman Test results for all evaluation metrics produced a p-value less than 0.05, indicating significant performance differences among the clustering methods. The t-SNE visualization and intuitive maps showed that province-level waste characteristics in Indonesia were not homogeneous and that several provinces experienced changes in cluster membership during the observation period. Overall, FRCM was the most competitive method for generating cluster structures based on the combination of the three primary evaluation metrics.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | K-Means, Fuzzy C-Means, Fuzzy Rough C-Means, Firefly Algorithm, data sampah, K-Means, Fuzzy C-Means, Fuzzy Rough C-Means, Firefly Algorithm, waste data. |
| Subjects: | Q Science > QA Mathematics > QA278.55 Cluster analysis Q Science > QA Mathematics > QA39.3 Fuzzy mathematics Q Science > QA Mathematics > QA9.64 Fuzzy logic Q Science > QA Mathematics > QA248_Fuzzy Sets |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis |
| Depositing User: | Raissa Undita Estiningtyas |
| Date Deposited: | 29 Jul 2026 03:08 |
| Last Modified: | 29 Jul 2026 03:08 |
| URI: | http://repository.its.ac.id/id/eprint/139212 |
Actions (login required)
 |
View Item |