A Comparative Study of Traditional, Ensemble and Neural Network-Based Natural Language Processing Algorithms
Bibliographic record
Abstract
Accurate data analysis is an important part of data-driven financial audits. Given the increased data availability and various systems from which audit files are generated, RCSFI provides a way for standardization on behalf of analysis. This research attempted to automate this hierarchical text classification task in order to save financial auditors time and avoid errors. Several studies have shown that ensemble-based models and neural-network-based natural language processing (NLP) techniques achieved encouraging results for classification problems in various domains. However, there has been limited empirical research comparing the performance of both of the aforementioned techniques in a hierarchical multi-class classification setting. Moreover, neural-network- based NLP techniques have commonly been applied to English datasets and not to Dutch financial datasets. Additionally, this research took the implementation of hierarchical approaches into account for the traditional and ensemble-based models and found that the performance did not increase when implementing the included hierarchical approaches. DistilBERT achieved the highest scores on level 1-2-3-4 and outperformed the traditional and ensemble-based models. The model obtained a F1 of 94.50% for level 1-2-3-4. DistilBERT also outperformed BERTje at level 1-2-3-4 despite BERTje being specifically pre-trained on Dutch datasets.
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 | 0.005 | 0.011 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.003 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.004 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
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".