Prediksi Profil Risiko Kepatuhan Wajib Pajak PBB-P2 di Kota Mojokerto Menggunakan Metode Convolutional Neural Network (CNN)

Wulangsih, D. Febry (2026) Prediksi Profil Risiko Kepatuhan Wajib Pajak PBB-P2 di Kota Mojokerto Menggunakan Metode Convolutional Neural Network (CNN). Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) 22082010188-cover.pdf Download (1MB)
	Text (bab 1) 22082010188-bab1.pdf Download (325kB)
	Text (bab 2) 22082010188-bab2.pdf Restricted to Repository staff only until 15 September 2028. Download (523kB)
	Text (bab 3) 22082010188-bab3.pdf Restricted to Repository staff only until 15 September 2028. Download (895kB)
	Text (bab 4) 22082010188-bab4.pdf Restricted to Repository staff only until 15 September 2028. Download (1MB)
	Text (bab 5) 22082010188-bab5.pdf Download (222kB)
	Text (daftar pustaka) 22082010188-daftarpustaka.pdf Download (178kB)
	Text (lampiran) 22082010188-lampiran.pdf Restricted to Repository staff only until 15 September 2028. Download (275kB)

Abstract

Taxpayer compliance of the Rural and Urban Land and Building Tax (PBB-P2) is an important aspect in supporting the optimization of regional tax revenues. Based on the exploration of payment data in 2020–2025, the distribution of payment status shows a tendency to decrease the number of tax objects with Paid status and an increase in the number of tax objects with Unpaid status. This study aims to build a Convolutional Neural Network (CNN) model in predicting PBB-P2 taxpayer compliance and forming a compliance risk profile in Mojokerto City in 2026. The 1D-CNN model was developed using payment history and tax object characteristics through testing nine hyperparameter scenarios, eight architectural variations, and four threshold variations. The data was divided using stratified random split with a proportion of 72% training, 8% validation, and 20% testing, with F1-Score as the main model selection metric. The results of the final model evaluation obtained an Accuracy of 0.8148, Precision 0.8652, Recall 0.7262, F1-Score 0.7896, and ROC-AUC of 0.8477 in the Unpaid class. The final model was applied to 1,751 NOP to produce a compliance probability in 2026 which was further interpreted into low, medium, and high risk profiles. The model was then implemented in a website-based interactive dashboard using Flask to present the results of compliance predictions and taxpayer risk profiles.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Permatasari, Reisa	NIDN0014059203	reisa.permatasari.sifo@upnjatim.ac.id
Thesis advisor	Mukhlis, Iqbal Ramadhani	NIDN0705039304	iqbal.ramadhani.fasilkom@upnjatim.ac.id

Subjects:

Q Science > Q Science (General)
Q Science > QA Mathematics > QA76.6 Computer Programming
T Technology > T Technology (General)

Divisions:

Faculty of Computer Science > Departemen of Information Systems

Depositing User:

D. Febry Wulangsih

Date Deposited:

15 Sep 2026 04:02

Last Modified:

15 Sep 2026 04:10

URI:

https://repository.upnjatim.ac.id/id/eprint/60240