datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
M-ABSA
M-ABSA
This repo contains the data for our paper M-ABSA: A Multilingual Dataset for Aspect-Based Sentiment Analysis.
Data Description:
This is a dataset suitable for the multilingual ABSA task with triplet extraction.
All datasets are stored in the data/ folder:
All dataset contains 7 domains.
domains = ["coursera", "hotel", "laptop", "restaurant", "phone", "sight", "food"]
Each dataset contains 21 languages.
langs = ["ar", "da", "de", "en", "es", "fr", "hi"… See the full description on the dataset page: https://huggingface.co/datasets/Multilingual-NLP/M-ABSA.en-fiqa-absa
Dataset Name
📄 Paper · 💻 Code · 🌐 The Fin AI
Used in FinBen — FinBen: A Holistic Financial Benchmark for Large Language Models (arXiv:2402.12659).
Dataset Description
This dataset is based on the task 1 of the Financial Sentiment Analysis in the Wild (FiQA) challenge. It follows the same settings as described in the paper 'A Baseline for Aspect-Based Sentiment Analysis in Financial Microblogs and News'. The dataset is split into three subsets: train, valid… See the full description on the dataset page: https://huggingface.co/datasets/TheFinAI/en-fiqa-absa.setfit-absa-semeval-restaurants
Dataset Card for "tomaarsen/setfit-absa-semeval-restaurants"
Dataset Summary
This dataset contains the manually annotated restaurant reviews from SemEval-2014 Task 4, in the format as
understood by SetFit ABSA.
For more details, see https://aclanthology.org/S14-2004/
Data Instances
An example of "train" looks as follows.
{"text": "But the staff was so horrible to us.", "span": "staff", "label": "negative", "ordinal": 0}
{"text": "To be completely fair, the only… See the full description on the dataset page: https://huggingface.co/datasets/tomaarsen/setfit-absa-semeval-restaurants.FiQA_ABSAsetfit-absa-semeval-laptops
Dataset Card for "tomaarsen/setfit-absa-semeval-laptops"
Dataset Summary
This dataset contains the manually annotated laptop reviews from SemEval-2014 Task 4, in the format as
understood by SetFit ABSA.
For more details, see https://aclanthology.org/S14-2004/
Data Instances
An example of "train" looks as follows.
{"text": "I charge it at night and skip taking the cord with me because of the good battery life.", "span": "cord", "label": "neutral", "ordinal": 0}… See the full description on the dataset page: https://huggingface.co/datasets/tomaarsen/setfit-absa-semeval-laptops.ABSADatasetssemeval-absaThis dataset is built as a playground for aspect-based sentiment analysis.absa-dataset
TRABL: Travel-Domain Aspect-Based Sentiment Analysis Dataset
This repository contains the TRABL dataset, released in support of our paper accepted to The ACM Web Conference 2026 (WWW 2026):
TRABL: A Unified Framework for Travel Domain Aspect-Based Sentiment Analysis Applications of Large Language Models
The dataset is designed to support research on Aspect-Based Sentiment Analysis (ABSA) in the travel domain, with a particular focus on joint extraction of structured sentiment… See the full description on the dataset page: https://huggingface.co/datasets/Booking-com/absa-dataset.VLSP2018-ABSA-Restaurant
VLSP2018-ABSA-Restaurant
Dataset Summary
The VLSP 2018 Restaurant corpus targets the same ACSA sub-tasks (ACD & SPC) on 4,751 Vietnamese restaurant reviews. This unified CSV includes:
12 aspect–category indicator columns, each with values {0, 1, 2, 3}.
A type column for train/dev/test.
A dataset column fixed to VLSP2018-ABSA-Restaurant.
Supported Tasks and Metrics
Aspect Category Detection
Sentiment Polarity Classification
Metrics: Precision, Recall, F1… See the full description on the dataset page: https://huggingface.co/datasets/visolex/VLSP2018-ABSA-Restaurant.VLSP2018-ABSA-Hotel
VLSP2018-ABSA-Hotel
Dataset Summary
The VLSP 2018 Hotel corpus is designed for Vietnamese Aspect-Based Sentiment Analysis (ABSA), covering two sub-tasks of Aspect Category Sentiment Analysis (ACSA):
Aspect Category Detection (ACD): identify which Aspect#Category pairs are present in each review.
Sentiment Polarity Classification (SPC): assign one of three sentiment labels (Positive, Negative, Neutral) to each detected Aspect#Category.
This unified CSV contains 5,600… See the full description on the dataset page: https://huggingface.co/datasets/visolex/VLSP2018-ABSA-Hotel.AbSanitas
Dataset Card for AbSanitas
Dataset summary
AbSanitas is a Spanish biomedical information retrieval dataset built from biomedical texts collected from official academic repositories and open-access sources.
This dataset is designed to support the training and evaluation of encoder models on biomedical retrieval and semantic matching tasks in Spanish.
Curated by: Barcelona Supercomputing Center (BSC)
Funded by: ALIA
Language(s) (NLP): Spanish (es)
License: CC BY-NC-ND 4.0… See the full description on the dataset page: https://huggingface.co/datasets/BSC-LT/AbSanitas.arab-absa-kp
arab-absa-kp — Arabic aspect-based sentiment analysis with key points
Arabic customer reviews (Jeeran) annotated for aspect-based sentiment analysis, where every
aspect span additionally carries the key points it expresses. One row per key point.
Each review is split into aspect spans; each span has its own sentiment and category; each span is
summarised as one or more key points, each with its own sentiment. A row therefore reads: in this
review, this span says this thing about… See the full description on the dataset page: https://huggingface.co/datasets/k-chirkunov/arab-absa-kp.Pars-ABSAsemeval2016-full-absa-reviews-english-translated-resampledABSA-alpaca-SemEval2014Task4semeval-2016-absa-reviews-arabic
Dataset Card for Dataset Name
Dataset Summary
Aspect based sentiment analysis dataset using hotel reviews in Arabic.
Languages
Arabic
Licensing Information
Original dataset was licensed under MIT, so this is also under MIT
Citation Information
Cite this and the original authors if you want to.
semeval-2016-absa-reviews-english-translated-stanford-alpaca
Dataset Card for Dataset Name
Derived from eastwind/semeval-2016-absa-reviews-arabic using Helsinki-NLP/opus-mt-tc-big-ar-en
semeval-2016-absa-reviews-english-translated-resampled
Dataset Card for Hotel Review ABSA (SemEval 2016 Translated from Arabic)
Dataset Description
Derived from eastwind/semeval-2016-absa-reviews-english-translated-stanford-alpaca, by upsampling the neutral class and then resampling 3k examples from each class
absa-quad
上述数据集为ABSA(Aspect-Based Sentiment Analysis)领域数据集，基本形式为从句子中抽取：方面术语、方面类别(术语类别)、术语在上下文中情感极性以及针对该术语的观点词，不同数据集抽取不同的信息，这点在jsonl文件的“instruction”键中有分别提到，在此我将其改造为了生成任务，需要模型按照一定格式生成抽取结果。
以acos数据集中抽取的jsonl文件一条数据举例：
{
"task_type": "generation",
"dataset": "acos",
"input": ["the computer has difficulty switching between tablet and computer ."],
"output": "[['computer', 'laptop usability', 'negative', 'difficulty']]",
"situation": "none",
"label": "",
"extra": ""… See the full description on the dataset page: https://huggingface.co/datasets/NEUDM/absa-quad.ABSA_beer
Unsupervised Aspect-Based Sentiment Analysis through LLM: A Case Study of an Unlabeled Portuguese Beer Database
This repository contains the datasets of the paper.
Code: https://github.com/deniseiras/ACADEMIC_ABSA_beer/releases/tag/v1.1.0
Datasets:
Reviews Main (step_3_reviews_main.csv):
Step 3 produced this dataset, which supports the analysis of review comments, quantitative beer attributes, and review-related information, as summarized in Table 2. From the 67,083 reviews… See the full description on the dataset page: https://huggingface.co/datasets/deniseiras/ABSA_beer.absa-tr
ABSA-TR
A Turkish aspect-based sentiment dataset with 16,031 real user-review sentences and 24,439 aspect annotations from e-commerce, supplements, and movie domains.
Data
Split
Sentences
Aspects
Implicit aspects
Train
11,993
17,423
3,995
Validation
688
1,218
179
Test
3,350
5,798
804
Each row contains pool_id, domain, text, aspects, and flags. Each aspect has a verbatim span, a normalized aspect, a polarity of positive, negative, or neutral… See the full description on the dataset page: https://huggingface.co/datasets/ytu-ce-cosmos/absa-tr.tubitak_clothing_absa_v3clothing_products_ABSA
Veri Kümesi Detayları
Veri Kümesinin Adı: ABSA (Aspect-Based Sentiment Analysis)
Size: 11470 veri (8029 train, 3441 test)
Language(s): Türkçe
Task: Aspect-Based Sentiment Analysis (ABSA)
Categories: Etek, Kaban, Gömlek, Kazak, Pantolon
Polarity: Negatif, Nötr, Pozitif
Aspect: Kalite, Kumaş, Renk, Beden, Kargo, Fiyat
License: MIT
Developed by: ebru kılıç , rumeysa nur yasav
Veri Kaynakları
Veriler Trendyol ve HepsiBurada sitelerinde bulunan giyim ürünlerine ait… See the full description on the dataset page: https://huggingface.co/datasets/ebrukilic/clothing_products_ABSA.clothing_products_ABSA_v2
Veri Kümesi Detayları
Veri Kümesinin Adı: ABSA (Aspect-Based Sentiment Analysis)
Size: 11470 veri (7460 train, 4010 test)
Language(s): Türkçe
Task: Aspect-Based Sentiment Analysis (ABSA)
Categories: Etek, Kaban, Gömlek, Kazak, Pantolon
Polarity: Negatif, Nötr, Pozitif
Aspect: Kalite, Kumaş, Renk, Beden, Kargo, Fiyat
License: MIT
Developed by: ebru kılıç , rumeysa nur yasav
Veri Kaynakları
Veriler Trendyol ve HepsiBurada sitelerinde bulunan giyim ürünlerine ait… See the full description on the dataset page: https://huggingface.co/datasets/ebrukilic/clothing_products_ABSA_v2.OATS-ABSA
OATS Dataset
Description
The OATS (Opinion Aspect Target Sentiment) dataset is a comprehensive collection designed for the Aspect Sentiment Quad Prediction (ASQP) or Aspect-Category-Opinion-Sentiment (ACOS) task. This dataset aims to facilitate research in aspect-based sentiment analysis by providing detailed opinion quadruples extracted from review texts. Additionally, for each review, we offer tuples summarizing the dominant sentiment polarity toward each aspect… See the full description on the dataset page: https://huggingface.co/datasets/jordiclive/OATS-ABSA.absa-semeval2014-alpacaExamples with one or more aspects that were labeled with the polarity 'conflict' were excluded. Examples are formatted in the Alpaca format. The purpose is to train an LLM top predict the aspects (output) based on the text (input).
@inproceedings{pontiki_semeval-2014_2014,
title = {{SemEval}-2014 {Task} 4: {Aspect} {Based} {Sentiment} {Analysis}},
doi = {10.3115/v1/S14-2004},
booktitle = {Proceedings of the 8th {International} {Workshop} on {Semantic} {Evaluation} ({SemEval} 2014)}… See the full description on the dataset page: https://huggingface.co/datasets/psimm/absa-semeval2014-alpaca.utmn-study-feedbacks-absa
Dataset Summary
A dataset for training and evaluating models in aspect-based sentiment analysis. It contains student reviews of academic courses written in Russian.
Dataset Structure
Data Fields
Input
text: review text
Output — sentiment labels for each aspect, the aspect is described in the column name:
лекции
доклады
проекты
презентации
фильмы
видео-уроки
задания__задачи
онлайн-курс
баллы__оценки
практики__семинары
тесты
домашняя работа
эссе
выступления… See the full description on the dataset page: https://huggingface.co/datasets/bulatovv/utmn-study-feedbacks-absa.semeval-2014-absasemeval-absaThis dataset is built as a playground for aspect-based sentiment analysis.absa-edu
Uzbek Education Aspect-Based Sentiment Analysis
Uzbek Education Aspect-Based Sentiment Analysis (absa-edu) is an aspect-level sentiment dataset containing Uzbek-language opinions about education. Each row provides a text sample, an aspect term and category, an opinion expression, sentiment and polarity labels, intensity, context, negation, irony, and script information.
Dataset Summary
Dataset ID: uznlp-uz/absa-edu
Language: Uzbek (uz)
Domain: Education
Rows: 7… See the full description on the dataset page: https://huggingface.co/datasets/uznlp-uz/absa-edu.