PENGEMBANGAN MODEL PENILAIAN KOMPLEKSITAS KASUS PEMERIKSAAN PAJAK DENGAN PENDEKATAN MACHINE LEARNING

Hidayatullah, Firda (2025) PENGEMBANGAN MODEL PENILAIAN KOMPLEKSITAS KASUS PEMERIKSAAN PAJAK DENGAN PENDEKATAN MACHINE LEARNING. Skripsi thesis, Politeknik Keuangan Negara STAN.

	Text (Cover) 01. Cover_Firda Hidayatullah_4132230059.pdf Restricted to Repository staff only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (210kB)
	Text (Abstrak) 02. Abstrak_Firda Hidayatullah_4132230059.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (193kB)
	Text (Daftar Isi) 03. Daftar Isi_Firda Hidayatullah_4132230059.pdf Restricted to Repository staff only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (303kB)
	Text (Bab I) 05. Bab I_Firda Hidayatullah_4132230059.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (443kB)
	Text (Bab II) 06. Bab II_Firda Hidayatullah_4132230059.pdf Restricted to Repository staff only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (489kB)
	Text (Bab III) 07. Bab III_Firda Hidayatullah.pdf Restricted to Repository staff only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (391kB)
	Text (Bab IV) 08. Bab IV_Firda Hidayatullah_4132230059.pdf Restricted to Repository staff only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (1MB)
	Text (Bab V) 09. Bab V_Firda Hidayatulah_4132230059.pdf Restricted to Repository staff only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (218kB)
	Text (Daftar Pustaka) 10. Daftar Pustaka_Firda Hidayatullah_4132230059.pdf Restricted to Repository staff only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (242kB)

Abstract

ABSTRAK Pemeriksaan pajak merupakan salah satu instrumen utama dalam memastikan kepatuhan wajib pajak serta optimalisasi penerimaan negara. Namun, proses distribusi kasus pemeriksaan pajak saat ini masih bersifat manual dan belum memiliki standar yang objektif dalam menilai kompleksitas suatu kasus. Ketidakseimbangan dalam distribusi kasus dapat menyebabkan beban kerja yang tidak proporsional serta menurunkan efektivitas pemeriksaan. Penelitian ini bertujuan untuk mengembangkan model penilaian kompleksitas kasus pemeriksaan pajak dengan pendekatan machine learning, guna meningkatkan objektivitas dan efisiensi dalam proses pemeriksaan. Metodologi yang digunakan dalam penelitian ini adalah Design Science Research Methodology (DSRM) yang terdiri dari enam tahapan: identifikasi masalah, perumusan tujuan solusi, desain dan pengembangan model, demonstrasi, evaluasi, serta komunikasi hasil penelitian. Model yang dikembangkan menggunakan pendekatan K-Means Clustering untuk mengelompokkan kasus pemeriksaan berdasarkan empat aspek utama kompleksitas: Quantity, Depend, Diverse, dan Change. Selain itu, metode Analytical Hierarchy Process (AHP) digunakan untuk memberikan bobot pada masing-masing aspek dalam proses klasifikasi. Data yang digunakan dalam penelitian ini berasal dari riwayat pemeriksaan pajak di Kantor Pelayanan Pajak (KPP) Pratama X selama periode 2015–2024. Hasil penelitian menunjukkan bahwa model berbasis machine learning yang dikembangkan mampu mengelompokkan kasus berdasarkan kompleksitas dengan lebih sistematis dibandingkan metode konvensional. Implementasi model ini dalam sistem informasi pemeriksaan pajak berpotensi meningkatkan akurasi dalam pendistribusian kasus serta mendukung alokasi sumber daya pemeriksa yang lebih efisien. Dengan adanya model ini, diharapkan distribusi pemeriksaan pajak menjadi lebih adil, meningkatkan efektivitas pemeriksaan, serta memperkuat transparansi dan akuntabilitas dalam proses perpajakan. Kata Kunci: Pemeriksaan Pajak, Kompleksitas Kasus, Machine Learning, K-Means Clustering, Analytical Hierarchy Process (AHP), Design Science Research Methodology (DSRM) xiv ABSTRACT Tax audits are a key instrument in ensuring taxpayer compliance and optimizing state revenue. However, the distribution of tax audit cases is still conducted manually and lacks an objective standard for assessing case complexity. The imbalance in case distribution can lead to an unequal workload and reduce audit effectiveness. Therefore, this study aims to develop a machine learning-based complexity assessment model for tax audit cases to enhance objectivity and efficiency in the auditing process. This study employs the Design Science Research Methodology (DSRM), which consists of six stages: problem identification, solution objective formulation, model design and development, demonstration, evaluation, and research findings communication. The proposed model utilizes K-Means Clustering to categorize audit cases based on four key complexity aspects: Quantity, Depend, Diverse, and Change. Additionally, the Analytical Hierarchy Process (AHP) method is applied to assign weights to each aspect in the classification process. The data used in this research comes from the tax audit history at Tax Office (KPP) Pratama X during the period 2015–2024. The findings indicate that the developed machine learning-based model can systematically classify cases by complexity more effectively than conventional methods. The implementation of this model within the tax audit information system has the potential to improve the accuracy of case distribution and support a more efficient allocation of tax auditors. By utilizing this model, tax audit distribution is expected to become fairer, enhance audit effectiveness, and strengthen transparency and accountability in the taxation process. Keywords: Tax Audit, Case Complexity, Machine Learning, K-Means Clustering, Analytical Hierarchy Process (AHP), Design Science Research Methodology (DSRM)

Actions (login required)

View Item