Dissertation Title: Forensic Accounting and the Detection of Fraud in Financial Statements 

Authors: Almina DOKO (MANOKU)

• Institution: University of Tirana,  Faculty of Economics, Department of Accounting

• Field of study: Accounting

• Publication date: 08.07.2026

• The dissertation is published in Albanian.

© Copyright: Almina DOKO (MANOKU)

Published by the University of Tirana
Based on legal acts, regulations and policies of the UT

👉 Click here to view the full dissertation (PDF)

 

Abstract

This paper treats forensic accounting as an integrated discipline that combines accounting, auditing, financial investigation and legal analysis for the detection and documentation of fraud in financial statements. International literature, in countries with developed forensic accounting practices, demonstrates the widespread use of statistical and algorithmic methods as an auxiliary instrument for the detection of financial fraud. In Albania and the Western Balkans, forensic accounting remains an unformalized discipline as an autonomous profession, and empirical studies that enrich it with artificial intelligence are lacking.

The study proposes a methodological framework for forensic accounting supported by artificial intelligence. The Beneish M-Score model is used for preliminary assessment of the risk of manipulation and classification of economic entities. This basis is used for training and testing of Support Vector Machines and Neural Networks algorithms, with the aim of developing predictive models that support the forensic accountant in identifying high-risk entities.

Empirical results show that integrating traditional financial analysis with artificial intelligence significantly increases the accuracy of fraud detection. Financial indicators with a crucial predictive role are identified, which can serve as early warning indicators. The study provides a viable framework for strengthening forensic accounting in Albania, with application in auditing, financial investigation and regulatory oversight.

Field: Accounting and Financial Analysis

Keywords: Financial Statement Fraud; Forensic Accounting; Beneish M-Score; Machine Learning

 

 

Abstract

This study addresses forensic accounting as an integrated discipline that combines accounting, auditing, financial investigation, and legal analysis for detecting and documenting financial statement fraud. International literature, particularly from countries with well-established forensic accounting practices, demonstrates the widespread use of statistical and algorithmic methods as supporting instruments for fraud detection. In Albania and the Western Balkans, forensic accounting remains unformalized as an autonomous profession, with a lack of empirical studies integrating it with artificial intelligence.

The study proposes a methodological framework for forensic accounting supported by artificial intelligence. The Beneish M-Score model is employed for preliminary assessment of manipulation risk and classification of economic entities. This foundation supports the training and testing of Support Vector Machines and Neural Networks algorithms, aiming to develop predictive models that assist the forensic accountant in identifying high-risk entities.

Empirical results indicate that integrating traditional financial analysis with artificial intelligence significantly enhances fraud detection accuracy. Financial indicators with decisive predictive roles are identified, which may serve as early warning signals. The study offers an applicable framework for strengthening forensic accounting in Albania, with applications in auditing, financial investigation, and regulatory oversight.

Field of study: Accounting and Financial Analysis

Keywords: Financial statement fraud; Forensic accounting; Beneish M-Score; Machine learning