adding to the clipboard ,order the refrences to what is mentioned first

Created using ChatSlide

This research proposes an Egyptian Arabic financial NLP framework combining five-class sentiment analysis of news and social media with market data. It addresses the limited availability of dialectal corpora, lexicons and benchmarks by developing an approximately 84,000-sample dataset and an AraBERT-based multi-modal pipeline. Drawing on regional evidence and hybrid architectures, the study will evaluate market-direction prediction, RMSE, error reduction and Granger causality, contributing...