Topic Modeling and Sentiment Analysis of Arabic News Headlines for a Societal Well-Being Scoring and Monitoring System: Moroccan Use Case

In today’s information-rich society, the rapid dissemination of news content across digital platforms plays a pivotal role in shaping public perception and discourse, yet they remain underutilized for sentiment analysis compared to other social media data. This study addresses this gap by using advanced Natural Language Processing (NLP) techniques to analyze Modern Standard Arabic (MSA) news content, categorizing headlines into themes like politics, business, education, weather, and sport, and applying sentiment analysis to classify them as positive or negative. These models form the foundation of our Societal Wellbeing Scoring and Monitoring System, which quantifies and monitors societal sentiment across key Quality of Life (QoL) aspects in Moroccan news. Our approach utilizes an LDA topic modeling approach in the preprocessing stage on the Moroccan News Arabic Dataset (MNAD) to enhance the quality of selected data sample before building classification models. Following this, we compare deep learning models with AraBERT and asafaya/bert-base-arabic (ASAFAYA) word embeddings to traditional machine learning algorithms using term frequency-inverse document frequency (TF-IDF). Notably, the GRU model with ASAFAYA embeddings excelled in the topic modeling task, achieving a 94.74% accuracy. We further analyzed public sentiment using data from Hibapress, a Moroccan news website, by scraping titles, likes, and dislikes to assign positive and negative sentiment scores. Leveraging a Bi-GRU model with ASAFAYA embeddings, we achieved an accuracy of 90.23%. We propose the Societal Wellbeing Scoring and Monitoring System, an aspect-based sentiment analysis framework designed to assign a wellbeing score over a specified sample or time period. Our findings highlight the effectiveness of combining topic modeling with deep learning-driven sentiment analysis to derive actionable insights from Arabic news, thereby enhancing the understanding of societal dynamics in Morocco.