BDNEWS: Advance Bengali Sensitivity Corpus and Classifiers
Bibliographic record
Abstract
The development of natural language processing has been facilitated by introducing numerous high-quality datasets and cutting-edge methodologies. Downstream NLP tasks, including sentiment analysis, are essential for comprehending the context and emotional condition of users. Nonetheless, a comprehensive understanding of the broader aspects of text context remains scarce. To address this gap, we introduce a large-scale dataset BDNEWS that encompasses three dimensions of sensitivity: Worldwide, Countrywise, and Personal which are further categorized into two states-Positive and Negative-resulting in a total of six distinct sentiment categories. We manually curated this dataset from various local Bengali online news portals, comprising 6,989 instances. Additionally, to obtain more effective contextual embeddings, we utilized unsupervised techniques such as FastText and Word2Vec, which were employed to train a bespoke Bidirectional Long Short-Term Memory (BiLSTM) model for precise sensitivity classification. FastText consistently surpasses Word2Vec in all metrics, illustrating its superiority in encapsulating the intricacies of Bengali language semantics. FastText achieves a maximum accuracy of 0.795 together with a recall score of 0.79 and a precision of 0.80. Moreover, its capacity to record sequential dependencies in text considerably improves the sensitivity prediction accuracy. These results give important new perspectives for sensitivity prediction in Bengali language applications and guide the next studies in this field.
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 | 0.001 | 0.006 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.007 | 0.006 |
| Science and technology studies | 0.002 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.022 | 0.030 |
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".