| --- |
| --- |
| language: |
| - en |
| - bem |
| - nya |
| license: apache-2.0 |
| base_model: google-bert/bert-base-multilingual-cased |
| tags: |
| - text-classification |
| - aspect-based-sentiment-analysis |
| - absa |
| - low-resource |
| - zambia |
| --- |
| |
| # baselineABSA |
|
|
| baselineABSA is a multilingual Aspect-Based Sentiment Analysis (ABSA) model |
| for Zambian ride-hailing social-media reviews. It is a generic multilingual |
| BERT model fine-tuned directly for aspect-conditioned sentiment |
| classification, with no prior localization. |
|
|
| ## Task |
|
|
| The model performs aspect-conditioned sentiment classification. Given a |
| review and a target service aspect, it predicts the sentiment expressed |
| toward that specific aspect, rather than an overall sentiment for the whole |
| review. |
|
|
| Sentiment classes: negative (0), neutral (1), positive (2). |
|
|
| Service aspects: driver_behavior, pricing, app_performance, payment, |
| ride_quality, customer_support, service_quality, booking, safety, |
| waiting_time. |
|
|
| ## Base model |
|
|
| google-bert/bert-base-multilingual-cased |
|
|
| ## Training data |
|
|
| The model was fine-tuned on a synthetic multilingual ride-hailing ABSA |
| dataset containing English, Bemba_Cibemba, Nyanja_Cinyanja, and Lusaka_Slang |
| reviews, with aspect-level sentiment annotations. |
| |
| ## Training procedure |
| |
| Standard fine-tuning of the full model. Training configuration: 7 epochs, |
| learning rate 2e-5, train and evaluation batch size 16, weight decay 0.01, |
| AdamW optimizer, evaluation and checkpoint saving per epoch, best checkpoint |
| selected on macro F1-score. |
| |
| ## Intended use |
| |
| baselineABSA was developed as part of a master's dissertation on |
| multilingual Aspect-Based Sentiment Analysis for low-resource Zambian |
| ride-hailing social-media discourse. It serves as the baseline system in a |
| comparative evaluation against ZambiaABSA, which uses an encoder adapted |
| through ZambiaSocialBERT. |
| |
| ## Limitations |
| |
| The model was trained on synthetic data; performance on naturally occurring |
| reviews may differ. Neutral sentiment and aspect entanglement remain |
| difficult, as documented in the associated dissertation. |
| --- |
| |
| |
|  |
| |
| |
|  |
| |
| |
|  |
| |
| |
| |
|  |
| |
| |
| |
|  |
| |
| |
| |
|  |
| |
| |
|  |
| |
| |
|  |
| |
| |
|  |
| |
| |
| |
|  |
| |
| |
| |
|  |
| |
| |
| |
| |
|  |
| |
| |
| |
|  |
| |
| |
| |
|  |
| |
| |
| |
|  |
| |
| |
| |
| |
| |
| |
| |
| |
| |
