Arabic language sentiment analysis on health services

Abstract

-The social media network phenomenon leads to a massive amount of valuable data that is available online and easy to access. Many users share images, videos, comments, reviews, news and opinions on different social networks sites, with Twitter being one of the most popular ones. Data collected from Twitter is highly unstructured, and extracting useful information from tweets is a challenging task. Twitter has a huge number of Arabic users who mostly post and write their tweets using the Arabic language. While there has been a lot of research on sentiment analysis in English, the amount of researches and datasets in Arabic language is limited. This paper introduces an Arabic language dataset which is about opinions on health services and has been collected from Twitter. The paper will first detail the process of collecting the data from Twitter and also the process of filtering, pre-processing and annotating the Arabic text in order to build a big sentiment analysis dataset in Arabic. Several Machine Learning algorithms (Naïve Bayes, Support Vector Machine and Logistic Regression) alongside Deep and Convolutional Neural Networks were utilized in our experiments of sentiment analysis on our health dataset.

A Combined CNN and LSTM Model for Arabic Sentiment Analysis

“…This is our own Arabic sentiment analysis dataset collected from Twitter. It was first presented in [9] and it has two classes (positive and negative). The dataset contains 2026 tweets and it is an unbalanced dataset that has 1398 negative tweets and 628 positive tweets.…”

Improving Sentiment Analysis in Arabic Using Word Representation

“…These approaches have increased the sentiment classification for our Arabic Health Services dataset (AHS) from 0.85 to 0.92 for the Main dataset, and from 0.87 to 0.95 for the Sub-dataset. Finally, this paper presents an improved accuracy, reaching 0.92, compared to our previous results in [1] that were 0.90 on the Main-Dataset.…”

“…In this experiment, the Main dataset is our previously proposed dataset of Arabic tweets about health services described in [1]. The dataset was collected from Twitter and contains 628 positive tweets, and 1398 negative tweets, to give a total of 2026.…”

“…As the dataset was labeled by only three human annotators, it can be hard to exactly confirm the positive or negative sentiment of each tweet, since sometimes the annotators disagreed. All the details on this can be found in [1]. We have extracted a subset from the main dataset, which we name the Sub-dataset.…”

“…Previous experiments on the main dataset were described in [1], where the accuracy results were between 0.85 and 0.90 using Naïve Bayes, Support Victor Machine, Logistic Regression and Basic Deep and Convolutional Neural Networks. In this experiment, different techniques will be used, focusing on employing different feature selection methods in order to improve the accuracy.…”