Abstract
Over the recent decades, there has been a significant increase and development of resources for Arabic natural language processing. This includes the task of exploring Arabic Language Sentiment Analysis (ALSA) from Arabic utterances in both Modern Standard Arabic (MSA) and different Arabic dialects. This study focuses on detecting sentiment in poems written in Misurata Arabic sub-dialect spoken in Misurata, Libya. The tools used to detect sentiment from the dataset are Sklearn as well as Mazajak sentiment tool1. Logistic Regression, Random Forest, Naive Bayes (NB), and Support Vector Machines (SVM) classifiers are used with Sklearn, while the Convolutional Neural Network (CNN) is implemented with Mazajak. The results show that the traditional classifiers score a higher level of accuracy as compared to Mazajak which is built on an algorithm that includes deep learning techniques. More research is suggested to analyze Arabic sub-dialect poetry in order to investigate the aspects that contribute to sentiments in these multi-line texts; for example, the use of figurative language such as metaphors.
Highlights
With the increasing, easier access to multiple Internet channels and social media platforms in which opinions about different topics and products are expressed, natural language processing (NLP) tools and services have shown a rapid improvement in order to investigate these structured and unstructured texts
By comparing the performance of AraBERT with that of BERT, the results indicate achieving a high level of accuracy with AraBERT on most of the assigned Arabic NLP tasks including Arabic Language Sentiment Analysis (ALSA)
Machine learning classifiers score a higher level of test accuracy than Mazajak which implements the Convolutional Neural Network (CNN) deep learning technique
Summary
Easier access to multiple Internet channels and social media platforms in which opinions about different topics and products are expressed, natural language processing (NLP) tools and services have shown a rapid improvement in order to investigate these structured and unstructured texts. Mining these utterances has led to the development of different applications and tasks. The process includes classifying extracted opinions in these texts into either objective or subjective text (Ghallab et al [40]). The subjective text can be classified into positive, negative, neutral, strongly negative, and strongly positive sentiments are included in some classifications
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have