Algorithmic Forensics: Detecting Earnings Management and Creative Accounting Practices at Meta Platforms, Inc. (2015–2024) Using Advanced Machine Learning Models

The increasing complexity of digital business models and the large-scale use of intangible assets within Big Tech firms present significant challenges to traditional auditing practices, potentially facilitating sophisticated forms of earnings management.This study addresses this gap by employing algorithmic forensics-a combination of established econometric models and advanced Machine Learning (ML) techniques-to identify and quantify potential creative accounting practices at Meta Platforms, Inc. (formerly Facebook) over the period 2015-2024.We calculate discretionary accruals using the Modified Jones Model as a baseline and then train various ML models (e.g., Random Forest, Gradient Boosting, LSTM networks) on a comprehensive dataset including financial statement data and non-financial metrics (e.g., Daily Active Users, Average Revenue Per User).Findings indicate that ML models significantly outperform the traditional econometric approach, achieving over 32% reduction in prediction error.Feature Importance Analysis Pinpoints R&D Intensity and Daily Active Users as the primary drivers of abnormal accruals.This research validates a novel framework for integrating ML into continuous auditing and offers evidence-based recommendations for audit regulators and practitioners operating in the digital economy.

• earnings management
• creative accounting
• Meta Platforms
• discretionary accruals
• modified Jones model
• continuous auditing
• Big Tech
• intangible assets
• R&D intensity
• feature importance
• LSTM networks
• gradient boosting
• random forest

• Auditing, Earnings Management, Governance
• Corporate Governance and Financial Management
• Benford’s Law and Fraud Detection
• earnings management
• creative accounting
• Meta Platforms
• discretionary accruals
• modified Jones model
• continuous auditing
• Big Tech
• intangible assets
• Accounting
• Strategy and Management
• Statistics and Probability
• Business, Management and Accounting
• Mathematics