Abstract:Sentiment analysis on social media platforms (i.e., Twitter or Facebook) has become an important tool to learn about users' opinions and preferences. However, the accuracy of sentiment analysis is disrupted by the challenges of natural language processing (NLP). Recently, deep learning models have proved superior performance over statistical- and lexical-based approaches in NLP-related tasks. Word embedding is an important layer of deep learning models to generate input features. Many word embedding models have been presented for text representation of both classic and context-based word embeddings. In this paper, we present a comparative analysis to evaluate both classic and contextualized word embeddings for sentiment analysis. The four most frequently used word embedding techniques were used in their trained and pre-trained versions. The selected embedding represents classical and contextualized techniques. Classical word embedding includes algorithms such as GloVe, Word2vec, and FastText. By contrast, ARBERT is used as a contextualized embedding model. Since word embedding is more typically employed as the input layer in deep networks, we used deep learning architectures BiLSTM and CNN for sentiment classification. To achieve these goals, the experiments were applied to a series of benchmark datasets: HARD, Khooli, AJGT, ArSAS, and ASTD. Finally, a comparative analysis was conducted on the results obtained for the experimented models. Our outcomes indicate that, generally, generated embedding by one technique achieves higher performance than its pretrained version for the same technique by around 0.28 to 1.8% accuracy, 0.33 to 2.17% precision, and 0.44 to 2% recall. Moreover, the contextualized transformer-based embedding model BERT achieved the highest performance in its pretrained and trained versions. Additionally, the results indicate that BiLSTM outperforms CNN by approximately 2% in 3 datasets, HARD, Khooli, and ArSAS, while CNN achieved around 2% higher performance in the smaller datasets, AJGT and ASTD.

Sentiment Analysis for Arabic Language Using Word Embedding

Improving Sentiment Analysis in Arabic Using Word Representation

Word Embeddings and Convolutional Neural Network for Arabic Sentiment Classification.

A Comparative Analysis of Word Embedding and Deep Learning for Arabic Sentiment Classification

Heterogeneous Ensemble Deep Learning Model for Enhanced Arabic Sentiment Analysis

Improving the Polarity of Text Through Word2vec Embedding for Primary Classical Arabic Sentiment Analysis

Natural Language Processing for Arabic Sentiment Analysis: A Systematic Literature Review

A Review for Arabic Sentiment Analysis Using Deep Learning

Subjective Sentiment Analysis for Arabic Newswire Comments

Sentiment Analysis For Modern Standard Arabic And Colloquial

Deep learning CNN–LSTM framework for Arabic sentiment analysis using textual information shared in social networks

Deep learning for emotion analysis in Arabic tweets

Aspect-based Sentiment Analysis and Location Detection for Arabic Language Tweets

Impact of Neural Network Architectures on Arabic Sentiment Analysis

Effect of Word Embedding Variable Parameters on Arabic Sentiment Analysis Performance

Emotion analysis of Arabic tweets using deep learning approach

Sentiment Analysis of Arabic Tweets: Feature Engineering and A Hybrid Approach

Sentiment analysis of Arabic social media texts: A machine learning approach to deciphering customer perceptions

Enhancing Arabic E-Commerce Review Sentiment Analysis Using a hybrid Deep Learning Model and FastText word embedding

An Enhanced Sentiment Analysis Framework Based on Pre-Trained Word Embedding

Improving Arabic sentiment analysis across context-aware attention deep model based on natural language processing