Sentiment Analysis of Iraqi Social Media Using Deep Learning: A TensorFlow Approach

Abstract

Sentiment analysis is a critical tool for understanding public opinion, especially in regions like Iraq where social media serves as a primary platform for expression. However, the complexities of the Iraqi Arabic dialect, including informal structures, slang, and spelling variations, pose significant challenges for traditional Natural Language Processing (NLP) models. This study addresses these challenges by developing a deep learning-based sentiment analysis model tailored specifically for Iraqi Arabic. Utilizing a dataset of 2,000 annotated tweets, we implement a Convolutional Neural Network (CNN) architecture optimized for binary sentiment classification. The model achieves an accuracy of 87.5% on the test set, demonstrating its effectiveness in capturing sentiment-bearing features in Iraqi dialect text. Key contributions include the creation of a labeled dataset for Iraqi Arabic and the development of a robust preprocessing and modeling pipeline. The findings highlight the potential of deep learning techniques for dialect-specific sentiment analysis, with applications in policymaking, business intelligence, and social media monitoring in Iraq and other Arabic-speaking regions.