Papers with Code · CC BY-SA 4.0

데이터셋 탐색

데이터셋의 설명·홈페이지를 찾고, 관련 태스크의 평가 기록을 살펴보세요.

관련 논문 87편

Asian Scientific Paper Excerpt Corpus

ASPEC
ASPEC 설명 보기

ASPEC, Asian Scientific Paper Excerpt Corpus, is constructed by the Japan Science and Technology Agency (JST) in collaboration with the National Institute of Information and Communications Technology (NICT). It consists of a Japanese-English paper abstract corpus of 3M parallel sentences (ASPEC-JE) and a Japanese-Chinese paper excerpt corpus of 680K parallel sentences (ASPEC-JC). This corpus is one of the achievements of the Japanese-Chinese machine translation project which was run in Japan from 2006 to 2010. Source: ASPEC Image Source: https://www.aclweb.org/anthology/L16-1350.pdf

관련 논문 64편

Aspect Sentiment Triplet Extraction

ASTE
ASTE 설명 보기

Target-based sentiment analysis or aspect-based sentiment analysis (ABSA) refers to addressing various sentiment analysis tasks at a fine-grained level, which includes but is not limited to aspect extraction, aspect sentiment classification, and opinion extraction. There exist many solvers of the above individual subtasks or a combination of two subtasks, and they can work together to tell a complete story, i.e. the discussed aspect, the sentiment on it, and the cause of the sentiment. However, no previous ABSA research tried to provide a complete solution in one shot. In this paper, we introduce a new subtask under ABSA, named aspect sentiment triplet extraction (ASTE). Particularly, a solver of this task needs to extract triplets (What, How, Why) from the inputs, which show WHAT the targeted aspects are, HOW their sentiment polarities are and WHY they have such polarities (i.e. opinion reasons). For instance, one triplet from “Waiters are very friendly and the pasta is simply average” could be (‘Waiters’, positive, ‘friendly’). We propose a two-stage framework to address this task. The first stage predicts what, how and why in a unified model, and then the second stage pairs up the predicted what (how) and why from the first stage to output triplets. In the experiments, our framework has set a benchmark performance in this novel triplet extraction task. Meanwhile, it outperforms a few strong baselines adapted from state-of-the-art related methods.

관련 논문 46편

Multi Aspect Multi-Sentiment

MAMS
MAMS 설명 보기

MAMS is a challenge dataset for aspect-based sentiment analysis (ABSA), in which each sentences contain at least two aspects with different sentiment polarities. MAMS dataset contains two versions: one for aspect-term sentiment analysis (ATSA) and one for aspect-category sentiment analysis (ACSA). Source: MAMS

관련 논문 11편

Aspect Category Opinion Sentiment

ACOS
ACOS 설명 보기

Most of the aspect based sentiment analysis research aims at identifying the sentiment polarities toward some explicit aspect terms while ignores implicit aspects in text. To capture both explicit and implicit aspects, we focus on aspect-category based sentiment analysis, which involves joint aspect category detection and category-oriented sentiment classification. However, currently only a few simple studies have focused on this problem. The shortcomings in the way they defined the task make their approaches difficult to effectively learn the inner-relations between categories and the inter-relations between categories and sentiments. In this work, we re-formalize the task as a category-sentiment hierarchy prediction problem, which contains a hierarchy output structure to first identify multiple aspect categories in a piece of text, and then predict the sentiment for each of the identified categories. Specifically, we propose a Hierarchical Graph Convolutional Network (Hier-GCN), where a lower-level GCN is to model the inner-relations among multiple categories, and the higher-level GCN is to capture the inter-relations between aspect categories and sentiments. Extensive evaluations demonstrate that our hierarchy output structure is superior over existing ones, and the Hier-GCN model can consistently achieve the best results on four benchmarks.

관련 논문 11편

Task Incremental Aspect Sentiment Classification

ASC (TIL, 19 tasks)
ASC (TIL, 19 tasks) 설명 보기

A set of 19 ASC datasets (reviews of 19 products) producing a sequence of 19 tasks. Each dataset represents a task. The datasets are from 4 sources: (1) HL5Domains (Hu and Liu, 2004) with reviews of 5 products; (2) Liu3Domains (Liu et al., 2015) with reviews of 3 products; (3) Ding9Domains (Ding et al., 2008) with reviews of 9 products; and (4) SemEval14 with reviews of 2 products - SemEval 2014 Task 4 for laptop and restaurant. For (1), (2) and (3), we split about 10% of the original data as the validate data, another about 10% of the original data as the testing data. For (4), We use 150 examples from the training set for validation. To be consistent with existing research(Tang et al., 2016), examples belonging to the conflicting polarity (both positive and negative sentiments are expressed about an aspect term) are not used. Statistics and details of the 19 datasets are given on Page https://github.com/ZixuanKe/PyContinual.

관련 논문 11편

Aspect Sentiment Quad Prediction

ASQP
ASQP 설명 보기

Aspect-based sentiment analysis (ABSA) typically focuses on extracting aspects and predicting their sentiments on individual sentences such as customer reviews. Recently, another kind of opinion sharing platform, namely question answering (QA) forum, has received increasing popularity, which accumulates a large number of user opinions towards various aspects. This motivates us to investigate the task of ABSA on QA forums (ABSA-QA), aiming to jointly detect the discussed aspects and their sentiment polarities for a given QA pair. Unlike review sentences, a QA pair is composed of two parallel sentences, which requires interaction modeling to align the aspect mentioned in the question and the associated opinion clues in the answer. To this end, we propose a model with a specific design of cross-sentence aspect-opinion interaction modeling to address this task. The proposed method is evaluated on three real-world datasets and the results show that our model outperforms several strong baselines adopted from related state-of-the-art models.

관련 논문 7편

Target Aspect Sentiment Detection

TASD
TASD 설명 보기

Aspect-based sentiment analysis (ABSA) aims to detect the targets (which are composed by continuous words), aspects and sentiment polarities in text. Published datasets from SemEval-2015 and SemEval-2016 reveal that a sentiment polarity depends on both the target and the aspect. However, most of the existing methods consider predicting sentiment polarities from either targets or aspects but not from both, thus they easily make wrong predictions on sentiment polarities. In particular, where the target is implicit, i.e., it does not appear in the given text, the methods predicting sentiment polarities from targets do not work. To tackle these limitations in ABSA, this paper proposes a novel method for target-aspect-sentiment joint detection. It relies on a pre-trained language model and can capture the dependence on both targets and aspects for sentiment prediction. Experimental results on the SemEval-2015 and SemEval-2016 restaurant datasets show that the proposed method achieves a high performance in detecting target-aspect-sentiment triples even for the implicit target cases; moreover, it even outperforms the state-of-the-art methods for those subtasks of target-aspect-sentiment detection that they are competent to.

관련 논문 4편

Conversational Aspect-based Sentiment Quadruple Extraction

DiaASQ
DiaASQ 설명 보기

DiaASQ is a fine-grained Aspect-based Sentiment Analysis (ABSA) benchmark under the conversation scenario. It challenges existing ABSA methods by 1) extracting quadruple of target-aspect-opinion-sentiment in a dialogue, and 2) modeling the dialogue discourse structures. The dataset is constructed by systematically crawling tweets from digital bloggers, followed by a series of preprocessing steps including filtering, normalizing, pruning, and annotating the collected dialogues, resulting in a final corpus of 1,000 dialogues. To enhance the multilingual usability, DiaASQ has both the English and Chinese versions of languages.

관련 논문 4편

MuSe-CarASTE: A comprehensive dataset for aspect sentiment triplet extraction in automotive review videos

MuseASTE
MuseASTE 설명 보기

•A new benchmark dataset for Aspect Sentiment Triplet Extraction. •First Aspect Sentiment Triplet Extraction (ASTE) Dataset in Automotive Domain. •Largest ASTE Dataset to date with annotations for over 28,295 sentences. •Dataset includes complex aspects not verbatim present in the sentence. •Domain: Aspect-based sentiment analysis, ASTE, Opinion Mining, Recommender System. •Four baseline SOTA models implemented on the dataset

관련 논문 2편

AWARE: Aspect-Based Sentiment Analysis Dataset of Apps Reviews for Requirements Elicitation

AWARE
AWARE 설명 보기

The peer-reviewed paper of AWARE dataset is published in ASEW 2021, and can be accessed through: http://doi.org/10.1109/ASEW52652.2021.00049. Kindly cite this paper when using AWARE dataset. Aspect-Based Sentiment Analysis (ABSA) aims to identify the opinion (sentiment) with respect to a specific aspect. Since there is a lack of smartphone apps reviews dataset that is annotated to support the ABSA task, we present AWARE: ABSA Warehouse of Apps REviews. AWARE contains apps reviews from three different domains (Productivity, Social Networking, and Games), as each domain has its distinct functionalities and audience. Each sentence is annotated with three labels, as follows: Aspect Term: a term that exists in the sentence and describes an aspect of the app that is expressed by the sentiment. A term value of “N/A” means that the term is not explicitly mentioned in the sentence. Aspect Category: one of the pre-defined set of domain-specific categories that represent an aspect of the app (e.g., security, usability, etc.). Sentiment: positive or negative. Note: games domain does not contain aspect terms. We provide a comprehensive dataset of 11323 sentences from the three domains, where each sentence is additionally annotated with a Boolean value indicating whether the sentence expresses a positive/negative opinion. In addition, we provide three separate datasets, one for each domain, containing only sentences that express opinions. The file named “AWARE_metadata.csv” contains a description of the dataset’s columns. How AWARE can be used? We designed AWARE such that it can be used to serve various tasks. The tasks can be, but are not limited to: Sentiment Analysis. Aspect Term Extraction. Aspect Category Classification. Aspect Sentiment Analysis. Explicit/Implicit Aspect Term Classification. Opinion/Not-Opinion Classification. Furthermore, researchers can experiment with and investigate the effects of different domains on users' feedback. Source: AWARE: Aspect-Based Sentiment Analysis Dataset of Apps Reviews for Requirements Elicitation

관련 논문 2편

Vietnamese Aspect-Based Sentiment Analysis Dataset

UIT-ViSFD
UIT-ViSFD 설명 보기

UIT-ViSFD is a Vietnamese Smartphone Feedback Dataset as a new benchmark corpus built based on strict annotation schemes for evaluating aspect-based sentiment analysis, consisting of 11,122 human-annotated comments for mobile e-commerce, which is freely available for research purposes.

관련 논문 1편

An aspect-based sentiment analysis dataset of Customer Feedback reviews

FABSA
FABSA 설명 보기

FABSA, An aspect-based sentiment analysis dataset in the Customer Feedback space (Trustpilot, Google Play and Apple Store reviews). A professionally annotated dataset released by Chattermill AI, with 8 years of experience in leveraging advanced ML analytics in the customer feedback space for high-profile clients such as Amazon and Uber. Two annotators possess extensive experience in developing human-labeled ABSA datasets for commercial companies, while the third annotator holds a PhD in computational linguistics. There has been a lack of high-quality ABSA datasets covering broad domains and addressing real-world applications. Academic progress has been confined to benchmarking on domain-specific, toy datasets such as restaurants and laptops, which are limited in size (e.g., SemEval Task ABSA or SentiHood). This dataset is part of the FABSA paper, and we release it hoping to advance academic progress as tools for ingesting and analyzing customer feedback at scale improve significantly, yet evaluation datasets continue to lag. FABSA is a new, large-scale, multi-domain ABSA dataset of feedback reviews, consisting of approximately 10,500 reviews spanning 10 domains (Fashion, Consulting, Travel Booking, Ride-hailing, Banking, Trading, Streaming, Price Comparison, Information Technology, and Groceries). Academic Paper @article{KONTONATSIOS2023126867, title = {FABSA: An aspect-based sentiment analysis dataset of user reviews}, journal = {Neurocomputing}, volume = {562}, pages = {126867}, year = {2023}, issn = {0925-2312}, doi = {https://doi.org/10.1016/j.neucom.2023.126867}, url = {https://www.sciencedirect.com/science/article/pii/S0925231223009906}, author = {Georgios Kontonatsios and Jordan Clive and Georgia Harrison and Thomas Metcalfe and Patrycja Sliwiak and Hassan Tahir and Aji Ghose}, keywords = {ABSA, Multi-domain dataset, Deep learning}, }

관련 논문 1편

Multi-aspect Integrated Migration Indicators dataset

MIMI dataset
MIMI dataset 설명 보기

Nowadays, new branches of research are proposing the use of non-traditional data sources for the study of migration trends in order to find an original methodology to answer open questions about cross-border human mobility. The Multi-aspect Integrated Migration Indicators (MIMI) dataset is a new dataset to be exploited in migration studies as a concrete example of this new approach. It includes both official data about bidirectional human migration (traditional flow and stock data) with multidisciplinary variables and original indicators, including economic, demographic, cultural and geographic indicators, together with the Facebook Social Connectedness Index (SCI). It results from the process of gathering, embedding and integrating traditional and novel variables, resulting in this new multidisciplinary dataset that could significantly contribute to nowcast/forecast bilateral migration trends and migration drivers. Thanks to this variety of knowledge, experts from several research fields (demographers, sociologists, economists) could exploit MIMI to investigate the trends in the various indicators, and the relationship among them. Moreover, it could be possible to develop complex models based on these data, able to assess human migration by evaluating related interdisciplinary drivers, as well as models able to nowcast and predict traditional migration indicators in accordance with original variables, such as the strength of social connectivity. Here, the SCI could have an important role. It measures the relative probability that two individuals across two countries are friends with each other on Facebook, therefore it could be employed as a proxy of social connections across borders, to be studied as a possible driver of migration. All in all, the motivations for building and releasing the MIMI dataset lie in the need of new perspectives, methods and analyses that can no longer prescind from taking into account a variety of new factors. The heterogeneous and multidimensional sets of data present in MIMI offer an all-encompassing overview of the characteristics of human migration, enabling a better understanding and an original potential exploration of the relationship between migration and non-traditional sources of data.

관련 논문 1편

Review level Opinion Aspect Sentiment Target Joint Detection for ABSA

ROAST
ROAST 설명 보기

This repository has a review-level multidomain multilingual dataset for Aspect-based Sentiment Analysis(ABSA) for the paper ROAST: Review-level Opinion Aspect Sentiment Target Joint Detection.

관련 논문 0편

Aspects dataset

Aspects dataset 설명 보기

This dataset contains video shots for two different classes: tigers and cars. The shots were collected from 188 car ads (~1–2 min each) and 14 nature documentaries about tigers (~40 min), amounting to roughly 14 h of video. The videos were partitioned into shorter shots, and only those showing at least one instance of the class were kept. This produced 806 shots for the car and 1880 for the tiger class, typically 1–100 sec in length. Paper: Discovering object aspects from video Source: Aspects dataset Image Source: Aspects dataset