Table 2. An overview of ASC benchmark datasets.

Where a represents aspect term; c represents aspect category; o represents opinion term; p represents sentiment polarity; the last column indicates the references associated with each dataset.

Dataset	Language	Annotations	References
SemEval-2014	English	a,c,p	Li et al. (2018), Chen et al. (2017), Gu et al. (2018), Ma et al. (2017), Wang, Huang & Zhao (2016), Tay, Tuan & Hui (2017), Chen, Teng & Zhang (2020), Li et al. (2021), Zhang, Li & Song (2019b), Xu et al. (2019), Pontiki et al. (2014), Zheng & Xia (2018), Brun, Popa & Roux (2014), Liu et al. (2018), Li et al. (2019), Tang et al. (2020), Kiritchenko et al. (2014), Majumder et al. (2018), Zeng, Ma & Zhou (2019), Nguyen & Shirai (2015)
SemEval-2015	English	a,c,p	Tay, Tuan & Hui (2017), Li et al. (2021), Zhang, Li & Song (2019b), Xu et al. (2019), Pontiki et al. (2015), Tang et al. (2020), Toh & Su (2016), Ma, Peng & Cambria (2018)
SemEval-2016	English	a,c,p	Tay, Tuan & Hui (2017), Zhang, Li & Song (2019b), Pontiki et al. (2016), Tang et al. (2020), Toh & Su (2016), Wang et al. (2016), Ruder, Ghaffari & Breslin (2016)
ASC-QA	Chinese	a,c,p	Wang et al. (2019b)
ASAP	Chinese	c,p	Bu et al. (2021)
ACOS	English	a,c,p,o	Cai, Xia & Yu (2021)
ABSA-QUAD	English	a,c,p,o	Zhang et al. (2021)
Twitter	English	a,c,p	Li et al. (2018), Li et al. (2021), Zhang, Li & Song (2019b), Bian et al. (2020), Tang et al. (2020), Liu & Zhang (2017), Tang et al. (2016),
Sentihood	English	a,c,p	Sun, Huang & Qiu (2019), Zhang, Zhang & Vo (2015)
Mitchell	English	a,c,p	Zhang, Zhang & Vo (2016), Liu & Zhang (2017), Marcheggiani et al. (2014)