Pengelompokan Perusahaan Asuransi Umum Berbasis Metode Fuzzy C-Means Dengan Optimasi Genetic Algorithm dan Particle Swarm Optimization

Zahiroh, Diana (2025) Pengelompokan Perusahaan Asuransi Umum Berbasis Metode Fuzzy C-Means Dengan Optimasi Genetic Algorithm dan Particle Swarm Optimization. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5006211024-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (2MB) | Request a copy

Abstract

Industri asuransi merupakan salah satu sektor keuangan nonperbankan yang berperan dalam menjaga stabilitas perekonomian nasional. Pertumbuhan perusahaan asuransi menuntut adanya evaluasi kinerja perusahaan asuransi untuk meminimumkan risiko fraud dan meningkatkan kepercayaan pemegang polis. Rasio keuangan merupakan faktor krusial dalam mitigasi risiko kebangkrutan dan pengambilan keputusan strategis, terutama pada perusahaan asuransi. Penelitian ini bertujuan untuk mengelompokkan perusahaan asuransi umum berdasarkan rasio keuangan menggunakan metode Fuzzy C-Means (FCM), serta membandingkan kinerja metode yang dioptimalkan, yaitu Genetic Algorithm-Fuzzy C-Means (GA-FCM) dan Particle Swarm Optimization-Fuzzy C-Means (PSO-FCM). Data yang digunakan dalam penelitian ini berasal dari laporan keuangan perusahaan asuransi umum tahunan 2023, dengan indikator meliputi rasio solvabilitas, rasio likuiditas, rasio beban, rasio kecukupan investasi, rasio perimbangan hasil investasi dengan pendapatan premi neto, cadangan klaim dan cadangan premi. Penelitian ini menggunakan metode clustering untuk mengelompokan perusahaan asuransi ke dalam beberapa cluster berdasarkan karakteristik keuangan perusahaan tersebut. Validasi hasil klasterisasi dilakukan menggunakan Modified Partition Coeffecient (MPC) dan Xie Beni Index (XBI) untuk menentukan hasil klaster paling optimal. Cluster yang terbentuk sebesar 2 cluster karena nilai MPC terbesar dari metode FCM clustering. Berdasarkan nilai MPC dan XBI secara berturut-turut, yaitu 0,639222 dan 0,608979, metode optimasi Particle Swarm Optimization-Fuzzy C-Means Clustering (PSO-FCM) memberikan hasil cluster lebih optimal dibandingkan dengan Fuzzy C-Means Clustering tanpa optimasi dan Fuzzy C-Means Clustering dengan optimasi Genetic Algorithm. Kelompok perusahaan asuransi umum pada cluster 1 beranggotakan 22 perusahaan dengan cluster 1 sebagai cluster perusahaan yang sehat berdasarkan rasio keuangan. Cluster 2 beranggotakan 42 perusahaan asuransi umum dengan cluster 2 sebagai cluster yang lebih perlu berhati-hati dalam pengambilan keputusan.
=======================================================================================================================================
The insurance industry is one of the non-bank financial sectors that plays a vital role in maintaining national economic stability. The growth of insurance companies necessitates performance evaluations to minimize the risk of fraud and enhance policyholder trust. Financial ratios are crucial factors in mitigating bankruptcy risk and supporting informed strategic decision-making, particularly in the insurance industry. This study aims to cluster general insurance companies based on financial ratios using the Fuzzy C-Means (FCM) method and to compare the performance of optimized clustering methods, specifically the Genetic Algorithm-Fuzzy C-Means (GA-FCM) and Particle Swarm Optimization-Fuzzy C-Means (PSO-FCM) methods. The data used in this study are derived from the 2023 annual financial reports of general insurance companies, with indicators including solvency ratio, liquidity ratio, expense ratio, investment adequacy ratio, ratio of investment returns to net premium income, claim reserves, and premium reserves. Clustering methods are employed to group the insurance companies based on their financial characteristics. The clustering results are validated using the Modified Partition Coefficient (MPC) and Xie-Beni Index (XBI) to determine the most optimal clustering outcome. Two clusters were formed, with the optimal number of clusters determined based on the highest MPC value from the FCM clustering method. Based on the MPC and XBI values is 0,639222 and 0,608979 respectively, the Particle Swarm Optimization-Fuzzy C-Means Clustering (PSO-FCM) method produced more optimal clustering results compared to both the standard Fuzzy C-Means and the Genetic Algorithm-optimized FCM. Cluster 1 consists of 22 general insurance companies categorized as financially healthy based on their financial ratios. Cluster 2 consists of 42 general insurance companies that require more caution in strategic decision-making.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Clustering, Fuzzy C-Means, Genetic Algorithm, Particle Swarm Optimization, Modified Partition Coefficient
Subjects:	Q Science > QA Mathematics > QA278.55 Cluster analysis Q Science > QA Mathematics > QA39.3 Fuzzy mathematics Q Science > QA Mathematics > QA402.5 Genetic algorithms. Interior-point methods. Q Science > QA Mathematics > QA9.64 Fuzzy logic Q Science > QA Mathematics > QA248_Fuzzy Sets Q Science > QA Mathematics > QA278 Cluster Analysis. Multivariate analysis. Correspondence analysis (Statistics)
Divisions:	Faculty of Mathematics, Computation, and Data Science > Actuaria > 94203-(S1) Undergraduate Thesis
Depositing User:	Diana Zahiroh
Date Deposited:	31 Jul 2025 04:24
Last Modified:	31 Jul 2025 04:24
URI:	http://repository.its.ac.id/id/eprint/124528

Actions (login required)

View Item