![](https://csdnimg.cn/release/download_crawler_static/78381881/bg4.jpg)
ABSA’s earliest efforts depended mainly on machine learning methods that
focus on handcrafted features like lexicons to train sentiment classifiers (Jiang
et al., 2011)(Kiritchenko et al., 2014). These methods are effective but rely
heavily on the efficiency of handcrafted features.
Subsequently, a set of methods based on neural networks consisting of a
word embedding layer followed by a neural architecture were developed for
the ABSA task and pretty results were achieved (Tang et al., 2015a; Zhang
et al., 2016; Poria et al., 2016).
Several attention-based models have been applied to the ABSA task for its
ability to focus on important parts of the sentence related to aspects (Wang
et al., 2016) (Yang et al., 2017). But a single attention network maybe not
enough at capturing key dependency attributes between context and targets
especially when the sentence is long, so multiple attention networks were
proposed to solve this problem (Chen et al., 2017; Ma et al., 2017).
In order to resolve the question-answering problem, the authors of (Weston
et al., 2014) have developed the memory network concept (MemNN), which
was later adopted in several NLP challenges including ABSA(Tang et al.,
2016; Tay et al., 2018, 2017).
Pre-trained language models have recently played a vital role in many NLP
applications, as they can take advantage of the vast volume of unlabeled data
to learn general language representations; Elmo (Peters et al., 2018), GPT
(Radford et al., 2018), and BERT (Devlin et al., 2018) are among the most
well-known examples. The authors of (Li et al., 2019) studied the use of
BERT embeddings with several neural models, such as a linear layer, GRU
model, self-attention network, and conditional random field to deal with the
ABSA task. In (Sun et al., 2019), the authors have demonstrated that treat-
ing ABSA as a sentence pair classification task by building auxiliary sentence
as input to the BERT, significantly improved the results. The strength of
BERT embeddings in ABSA was further investigated by (Hoang et al., 2019;
Bhoi and Joshi, 2018; Xu et al., 2019; Nurifan et al., 2019).
In general, Arabic ABSA research development is slower than English.
(Ruder et al., 2016b) combined aspect embedding with each word embedding
and sent the mixture to CNN model to address aspect polarity and category
identification tasks. Subsequently, (Ruder et al., 2016a) took advantage of
modeling internal information related to the hierarchical review structure in
solving the ABSA task.
Two supervised machine learning-based techniques, RNN and SVM supple-
mented with a set of handcrafted features, were suggested by (Al-Smadi et al.,
4