Abstract: Analyzing and classifying customer reviews in Arabic presents a significant challenge due to the diverse nature of the Arabic language, encompassing various dialects and nuances. In this paper, we address this challenge by employing machine learning and deep learning techniques to analyze and classify Arabic customer reviews. Our study is based on a comprehensive dataset compiled from various sources, containing 33,333 positive and 33,333 negative reviews after filtering out mixed sentiments. To tackle this task, we explore a range of machine learning algorithms, including Support Vector Machines (SVM), K-Nearest Neighbors (KNN), Random Forest, Decision Trees, and Logistic Regression, along with various feature extraction methods such as TF-IDF and Word2Vec. Additionally, we delve into the realm of deep learning by employing Convolutional Neural Networks (CNN), Long Short-Term Memory networks (LSTM), and Bidirectional LSTM (BiLSTM) to further enhance sentiment analysis performance. Our experiments in ML reveal that SVM with an RBF kernel and W2V and Ngrams feature extraction achieves the highest F1-score of 82.5% . Among the DL models, BiLSTM with 128 units and dropout = 0.2 emerges as the top performer with an F1-score of 87%. These findings underscore the effectiveness of deep learning techniques in handling the complexities of Arabic text analysis. our research provides valuable insights into sentiment analysis of Arabic customer reviews and presents a comprehensive evaluation of machine learning and deep learning algorithms, paving the way for enhanced customer feedback analysis in the Arabic-speaking world.

sentiment analysis, Natural Language Processing, Arabic reviews Analysis, Arabic Language Processing, Deep learning, Machine Learning.