Authorship Attribution in Bangla Literature: A Comparative Study of Word Embedding and Deep Learning Models
Bibliographic record
Abstract
Authorship attribution in Bangla Literature is an inherently difficult problem because of the language’s morphology, diglossia, and lack of sufficient computational resources. Filling this void, this paper explores embeddingbased neural methods to encode stylistic and semantic marks of Bangla writers. A large text corpus consisting of stories, novels, and essays by twenty renowned authors was collected, preprocessed with a pipeline centered around Bangla, and uniformly segmented into passages to ensure consistent modeling. Several word embeddings, such as FastText and Word2Vec with CBOW and Skip-gram training methods, were combined with deep learning classifiers such as CNN, LSTM, MLP, Bi-LSTM, and FastText with Hierarchical Softmax. The experimental results show that Skip-gram embeddings outperformed CBOW in all aspects of contextual dependencies capture, with Word2Vec Skip-gram and CNN classifier combination achieving the best accuracy rate of 98.54%. While both Bi-LSTM and Hierarchical Softmax models provided relatively poor performance, they still illustrate valuable contributions to sequence modeling and inexpensive classification. This study indicates the framework can be used for sensible structured preprocessing and neural models based on embeddings to assess authorship attribution for Bangla literature, while adding a consistent structure between text representation and classification. Future directions will explore extending it to more authors, genres, and historical writings, while applying cross-dataset evaluation and adversarial examples to study robustness. This study builds on the knowledge base for advancing computational authorship studies in Bangla literature, while also paving the way for similar studies in other morphologically rich resource-deficient languages.
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 | 0.002 | 0.007 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.002 | 0.003 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| 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".