Pemodelan klaim asuransi kendaraan bermotor dengan regularisasi pada artificial neural network = Modeling motor vehicle insurance’s claims with regularization in artificial neural network

Wijaya, Darmawan Putra (2025) Pemodelan klaim asuransi kendaraan bermotor dengan regularisasi pada artificial neural network = Modeling motor vehicle insurance’s claims with regularization in artificial neural network. Bachelor thesis, Universitas Pelita Harapan.

[thumbnail of Title] Text (Title)
Title.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (252kB)
[thumbnail of Abstract] Text (Abstract)
Abstract.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (495kB)
[thumbnail of Chapter 1] Text (Chapter 1)
Chapter 1.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (504kB)
[thumbnail of Chapter 2] Text (Chapter 2)
Chapter 2.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (2MB)
[thumbnail of Chapter 3] Text (Chapter 3)
Chapter 3.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (668kB)
[thumbnail of Chapter 4] Text (Chapter 4)
Chapter 4.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (2MB)
[thumbnail of Chapter 5] Text (Chapter 5)
Chapter 5.pdf
Restricted to Registered users only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (472kB)
[thumbnail of Bibliography] Text (Bibliography)
Bibliography.pdf
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (475kB)
[thumbnail of Appendices] Text (Appendices)
Appendices.pdf
Restricted to Repository staff only
Available under License Creative Commons Attribution Non-commercial Share Alike.

Download (2MB)

Abstract

Metode konvensional yang digunakan dalam memodelkan klaim asuransi adalah Generalized 'Linear Model (GLM). Akan tetapi, GLM memerlukan asumsi distribusi. Dengan berkembangnya Artificial Neural Network (ANN), masalah pemodelan dapat diselesaikan tanpa mengasumsikan distribusi. Pada penelitian ini, akan diimplementasikan model ANN dalam memodelkan klaim asuransi kendaraan bermotor. Model ANN yang digunakan memiliki fleksibilitas pada nilai hyperparameter. Setelah model diselesaikan hyperparameter model ANN ditetapkan, parameter dengan mengimplementasi algoritma < untuk menyelesaikan model dengan nilai Mean Square Error (MSE) terkecil. Pemilihan hyperparameter Adam dilakukan Network (EvoNN). dengan Algoritma menggunakan algoritma Evolving Neural EvoNN memodifikasi model ANN menjadi beberapa model turunan (offspring). Model-model turunan selanjutnya diseleksi berdasarkan nilai Absolute Percentage Overall Error Sebagai pembanding, dibuat model GLM yang ditransformasi ke dalam model ANN (APOE). (GLMNN). Pemilihan parameter = model GLMNN dilakukan dengan mengimplementasikan algoritma Adam untuk menyelesaikan model dengan nilai Negative Log Likelihood Estimator (NLLE) terkecil. "Model GLMNN dibuat dengan mengasumsikan data berdistribusi Negatif Binomial. Hasil penelitian menunjukkan bahwa algoritma EvoNN dapat dengan efektif mengoptimalkan model ANN. Model yang optimal umumnya memiliki nilai iyperparameter yang besar dengan suatu batas atas dan menggunakan fungsi aktifasi tanh dan ELU. Selain itu, model yang paling optimal dalam meninimalkan nilai APOE belum tentu memiliki nilai MSE terkecil. Terakhir, model GLM dapat dipandang sebagai sebuah model GLMNN sehingga dapat diselesaikan secara numerik, namun proses penyelesaian model masih belum optimal sehingga model yang didapat kurang akurat dalam memprediksi klaim. / Conventional method often used to model insurance claims includes Generalized Linear Model (GLM). However, GLM requires the assumption of distribution. With the development of Artificial Neural Networks (ANN), modeling problems can be solved without assuming any distribution. In this research, ANN will be implemented to model auto insurance claims. The ANN model accommodates different selected, hyperparameters value. the Once the hyperparameter architecture are model parameters can be optimized using Adam’s algorithm to produce the least Mean Square Error (MSE) Value. The hyperparameter tuning will be done using the Evolving Neural Network (EvoNN) algorithm by modifying model onto several derivatives of the ANN model (offspring) and selecting the best based of their Absolute Percentage Overall Error (APOE). As a comparison, GLM with be transformed onto ANN (GLMNN). To solve the GLMNN model’s parameter, Adam’s algorithm will be used to minimize the value of the Negative Log Likelihood Estimator (NLLE). The GLMNN model used will assume that the data follows a Negative Binomial distribution. algorithm effectively optimizes ANN model. The The results show that EvoNN optimal model generally has large hyperparameter values within a certain upper limit and uses tanh and ELU activation functions. Additionally, the model that minimizes APOE doesn’t necessarily achieve the lowest MSE value. Lastly, GLM can transformed onto GLMNN and solved numerically. However, the model hasn’s been optimized properly causing less accurate claim predictions.
Item Type: Thesis (Bachelor)
Creators:
Creators
NIM
Email
ORCID
Wijaya, Darmawan Putra
NIM01112200010
darmawan88.putrawijaya@gmail.com
UNSPECIFIED
Contributors:
Contribution
Contributors
NIDN/NIDK
Email
Thesis advisor
Saputra, Kie Van Ivanky
NIDN0401038203
kie.saputra@uph.edu
Thesis advisor
Lukas, Samuel
NIDN0331076001
samuel.lukas@uph.edu
Uncontrolled Keywords: artificial neural network; evolving neural network; generalized linear model; regularisasi; algoritma adam; optimasi; klaim kendaraan bermotor; artificial neural network; evolving neural network; generalized linear model; regularization; adam’s algorithm; optimization; motor vehicle’s claims.
Subjects: Q Science > QA Mathematics
Divisions: University Subject > Current > Faculty/School - UPH Karawaci > Faculty of Science and Technology > Mathematics
Current > Faculty/School - UPH Karawaci > Faculty of Science and Technology > Mathematics
Depositing User: Stefanus Tanjung
Date Deposited: 09 Aug 2025 06:02
Last Modified: 09 Aug 2025 06:02
URI: http://repository.uph.edu/id/eprint/70422

Actions (login required)

View Item
View Item