---
task_categories:
- text-generation
language:
- en
---
> 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。

补充：SemEval-2014数据集文件夹中有两个文件夹"laptop"和"restaurant"，其实根据数据集文本的主要围绕主题区分的。抽取的元素方面，laptop和restaurant两文件夹中，数据的抽取元素也不同，laptop抽取的是方面类别和情感极性、restaurant抽取的是{(方面术语，情感极性)，(方面类别，情感极性)}的元素

#### 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": "", 
    "instruction": "    
        Task: Extracting aspect terms and their corresponding aspect categories, sentiment polarities, and opinion words. 
        Input: A sentence
        Output: A list of 4-tuples, where each tuple contains the extracted aspect term, its aspect category, sentiment polarity, and opinion words (if any). Supplement: \"Null\" means that there is no occurrence in the sentence.
        Example:  
            Sentence: \"Also it's not a true SSD drive in there but eMMC, which makes a difference.\"  
            Output: [['SSD drive', 'hard_disc operation_performance', 'negative', 'NULL']]' 
    "
}
```
> 此处未设置label和extra，在instruction中以如上所示的字符串模板，并给出一个例子进行one-shot，ABSA领域数据集(absa-quad,acos,arts,aste-data-v2,mams,semeval-2014,semeval-2015,semeval-2016,towe)每个数据集对应instruction模板相同，内容有细微不同，且部分数据集存在同一数据集不同数据instruction内容不同的情况。


#### 原始数据集
- 数据[链接](https://alt.qcri.org/semeval2014/task4/)
- Paper：[SemEval-2014 Task 4: Aspect Based Sentiment Analysis](https://aclanthology.org/S14-2004/)
- 说明：数据分为Laptop和restaurant两个主题的数据，分别在两个文件夹中放置。两个主题的数据抽取的元素不同。


#### 当前SOTA
*数据来自[PaperWithCode](https://paperswithcode.com/sota)*

- [SemEval2014-Laptop](https://paperswithcode.com/sota/aspect-based-sentiment-analysis-on-semeval-5)
    - 评价指标：F1-score
    - 模型：InstructABSA (**79.34**)
    - Paper：[InstructABSA: Instruction Learning for Aspect Based Sentiment Analysis](https://paperswithcode.com/paper/instructabsa-instruction-learning-for-aspect)

- [SemEval2014-Restaurant](https://paperswithcode.com/sota/aspect-based-sentiment-analysis-on-semeval-5)
    - 评价指标：Accuracy(抽取的分类准确率)
    - 模型：HGCN (**84.09**)
    - Paper：[Learn from Structural Scope: Improving Aspect-Level Sentiment Analysis with Hybrid Graph Convolutional Networks](https://paperswithcode.com/paper/learn-from-structural-scope-improving-aspect)