An Efficient Resampling Technique for Financial Statements Fraud Detection: A Comparative Study
Bibliographic record
Abstract
Financial statement fraud detection is the process of identifying falsified financial statements. Traditional auditing methods are time-consuming, expensive, and subject to error. Therefore, adopting an efficient and robust machine learning mechanism is important. Unfortunately, the current data sources suffer from a severe class imbalance. The lack of sufficient fraudulent financial statement records inspires the use of various resampling techniques. This paper a) examines the efficiency of different resampling strategies to detect fraudulent financial statements while employing multi-layer feedforward neural networks, support vector machines, and naïve Bayes machine learning models, and b) investigates the superiority of using Raw Accounting Variables (RAVs) over financial ratios for financial statement fraud detection. A benchmark dataset of numerical financial variables (RAVs and financial ratios) is used as features for model evaluation. The fraud labels correspond to the Accounting and Auditing Enforcement Releases by the U.S. Securities and Exchange Commission (SEC). We analyze the performance of the models on 28 RAVs and 14 financial ratios suggested by accounting experts. Using the area under the receiver operating characteristic curve (AUC) as the performance metric, the synthetic minority oversampling technique (SMOTE), along with a three-layer feedforward neural network (AUC: 0.863), greatly outperformed the RUSBoost (AUC: 0.717) model.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.008 | 0.018 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".