Arabic Corpus for Figurative Sentiment Analysis

Abstract

Annotation texts for the Aspect Level Sentiment Analysis (ABSA) of the Arabic language is a necessary but time-consuming and costly method. The annotation process is needed to classify collected texts into figurative devices, such as hyperboles or similes. In this paper, we present a Figurative Corpus for Arabic Language (FCAL) approach for the annotated figurative sentiment corpus for Arabic texts which have been collected from the Holy Quran and Al-hadeeth as an MSA (Modern Standard Arabic). Accordingly, in this paper, we propose a model that can recognize and categorize the Arabic figurative devices automatically and conclude on the essential features through MSA texts. In addition, these would be based on supervised machine learning techniques.