FRAUD LENS: SEE BEYOND THE NUMBERS

About the Project

FraudLens: See Beyond the Numbers is a text-based corporate fraud risk intelligence platform developed to identify early narrative signals associated with the likelihood of fraudulent financial reporting. Traditional fraud detection approaches often rely on financial ratios and structured indicators, which may overlook important behavioural and communication signals embedded in corporate disclosures. FraudLens addresses this gap by analysing Management Discussion and Analysis (MD&A) narratives from annual reports using advanced textual analysis techniques. The platform integrates Sentiment Analysis and Topic Modelling using Latent Dirichlet Allocation (LDA) to examine management tone, key disclosure themes and potential inconsistencies between narrative sentiment and underlying topics. This combination introduces a Narrative Dissonance perspective, where positive or persuasive disclosure tone is examined alongside risk-related themes such as loss, restructuring, risk management and non-performing activities. The textual signals are further integrated with corporate governance and ownership structure factors to provide a broader assessment of corporate fraud risk. FraudLens translates these analytical results into a Fraud Risk Score and company risk ranking, supported by key risk indicators, trends and explanatory insights. The platform is designed to support regulators, auditors, investors, financial institutions, boards and audit committees in conducting more focused risk assessment and further investigation. By transforming unstructured corporate narratives into actionable fraud-risk intelligence, FraudLens provides a scalable approach to complement conventional financial-based fraud detection and enables stakeholders to see beyond the numbers.