A Hybrid Deep Learning and Stylometric Feature-Based Framework for Authorship Attribution in Bangla Literature
Bibliographic record
Abstract
Attribution of authorship is the act of studying authors' writing styles to determine the author of the document from a set of potential authors. The attribution of authorship in Bangla literature has unique challenges, including the rich morphology of Bangla, its diglossia, and limited annotated corpora. This study addresses multi-dimensional feature extraction and classification of authorship identification of 20 prominent Bangla authors. This research introduces a novel dataset called “Authorship Attribution Bangla Dataset 20,” which includes 20 prominent writers, over 15 million words, and more than 20,000 samples. The study follows a tri-level feature framework using lexical, syntactic, and semantic dimensions with stylometric features, TF-IDF representations, and advanced similarity approaches. Classical machine learning models were evaluated, including SVM, Random Forest, and MLP, and transformer-based models, namely BanglaBERT and multilingual BERT (mBERT). The results show that BanglaBERT outperforms the remaining models with 99.09% accuracy, demonstrating how pre-trained language models are useful in detecting complex stylistic and semantic patterns. We also conducted some adversarial authorship experiments that verified the robustness of our approach. In this study, the term 'hybrid' denotes a two-parallel-branch framework that integrates both deep learning (BanglaBERT, mBERT) and stylometric feature-based models, which are trained and evaluated separately within the same pipeline. In conclusion, we find that traditional approaches to stylometric analysis complement contemporary approaches that tokenize in their encapsulated, contextual embeddings, making it a proven, reliable, and high-precision framework for Bangla authorship attribution. The research contributes to the wider Bangla authorship attribution literature and provides a framework for future.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.004 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.002 | 0.008 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.003 |
| Insufficient payload (model declined to judge) | 0.000 | 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 teacher head, 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".