Semantic Partitioning and Machine Learning in Sentiment Analysis

The outbreak of coronavirus disease (COVID-19) has affected almost all of the countries of the world, and has had significant social and psychological effects on the population. Nowadays, social media platforms are being used for emotional self-expression towards current events, including the COVID-19 pandemic. The study of people’s emotions in social media is vital to understand the effect of this pandemic on mental health, in order to protect societies. This work aims to investigate to what extent deep learning models can assist in understanding society’s attitude in social media toward COVID-19 pandemic. We employ two transformer-based models for fine-grained sentiment detection of Arabic tweets, considering that more than one emotion can co-exist in the same tweet. We also show how the textual representation of emojis can boost the performance of sentiment analysis. In addition, we propose a dynamically weighted loss function (DWLF) to handle the issue of imbalanced datasets. The proposed approach has been evaluated on two datasets and the attained results demonstrate that the proposed BERT-based models with emojis replacement and DWLF technique can improve the sentiment detection of multi-dialect Arabic tweets with an F1-Micro score of 0.72.

Social media has become a vital platform for individuals, organizations, and governments worldwide to communicate and express their views. During the coronavirus disease 2019 (COVID-19) pandemic, social media sites play a crucial role in people communicating, sharing, and expressing their perceptions on various topics. Analyzing such textual data can improve the response time of governments and organizations to act on alarming issues. This study aims to perform sentiment analysis on the subject of COVID-19 vaccination, perform temporal and spatial analyses of the textual data, and find the most frequently discussed topics that may help organizations bring awareness to those topics. In this work, the sentiment analysis of tweets was performed using 14 different machine learning classifiers and natural language processing (NLP). Lexicon-based TextBlob and Vader are used for annotating the data. A natural language toolkit is used for preprocessing of textual data. Our analysis observed that unigram models outperform bigram and trigram models for all four datasets. Models using term frequency-inverse document frequency (TF-IDF) have higher accuracy than models using count vectorizer. In the count vectorizer class, logistic regression has the best average accuracy with 91.925%. In the TF-IDF class, logistic regression has the best average accuracy of 92%; logistic regression has the highest average recall, F1-score, and ten cross-validation scores, and a ridge classifier has the highest average precision. The unigram models show a standard deviation (SD) of less than 1 for all classifiers except for the Gaussian Naïve Bayes showing 1.18. The experimental results reveal the dates and times in which most positive, negative, and neutral tweets are posted.

In this paper, we describe the work developed in the creation of syllable-based language model for continuous speech recognition system for Myanmar language. Speech recognition systems contain language model as the part of prediction word order sequence. In English and other languages, speech recognition system can use word-based language model. Since Myanmar is monosyllabic and syllable-timed language, the syllable-based language model is more suitable in speech recognition system. At the first step, this paper explains the structure of traditional (Phrase based) language model and describes how to normalized Myanmar sentences and building of normalized Myanmar language model. In the second step, the evaluation of the two language models is expressed. Finally, the comparison of these two language models with the perplexity values is described. The syllable-based Myanmar language model has lower perplexity value than the traditional language model. This language model can support the creation of word-based Myanmar language model.

Despite the superior performance of large language models to generate natural language texts, it is hard to generate texts with correct logic according to a given task, due to the difficulties for neural models to capture strict logic from free-form texts. In this paper, we propose a novel graph-based language model, Logical-GLM, to extract strict logic from free-form texts and then infuse into language models. Specifically, we first capture information from natural language instructions and construct logical probability graphs that generally describe domains. Next, we generate logical skeletons to guide language model training, infusing domain knowledge into language models. At last, we alternately optimize the searching policy of graphs and language models until convergence. The experimental results show that Logical-GLM is both effective and efficient compared with traditional language models, despite using smaller-scale training data and fewer parameters. Our approach can generate instructional texts with more correct logic owing to the internalized domain knowledge. Moreover, the search of logical graphs reflects the inner mechanism of the language models, which improves the interpretability of black-box models.

<span lang="EN-US">Nowadays, predicting the success of a new movie is a crucial task. In this work, the hybrid approach considers the movie features as well as sentiment expressed in the movie review to predict the success rate of a movie. Multiple movie features such as title, director, star cast, and writer. Are considered for prediction. The related raw data is collected from the internet movie database (IMDb) website and after pre-processing, the collected data is used to generate the supervised machine learning model. Different supervised learning models are compared and the one with the best results is used further. The mean squared error, root mean squared error and r2 score of the models generated are comparable with existing models. Further, sentiment analysis of the movie-related tweets is performed. The accuracy of best sentiment analysis model is 88.47%. Finally, the two models are combined to give the success prediction rating of new movies and the results of the hybrid model are encouraging. The proposed model may be used to find the top-rated movies of a particular calendar year.</span>

Social networks are popular for advertising, idea sharing, and opinion formation. Due to COVID-19, coronavirus information disseminated on social media affects people’s lives directly. Individuals sometimes managed it well, but it often hampered daily activities. As a result, analyzing people’s feelings is important. Sentiment analysis identifies opinions or sentiments from text. In this paper, we present an effective model that leverages the benefits of Convolutional Neural Network (CNN) and Bidirectional Long Short-Term Memory (BiLSTM) to categorize Arabic tweets using a stacked ensemble learning model. First, the tweets are represented as vectors using a word embedding model, then the text feature is extracted by CNN, and finally the context information of the text is acquired by BiLSTM. Aravec, FastText, and ArWordVec are employed separately to assess the impact of the word embedding on the our model. We also compare the proposed method to various deep learning models: CNN, LSTM, and BiLSTM. Experiments are performed on three different Arabic datasets related to COVID-19 and vaccines. Empirical findings show that the proposed model outperformed the other models’ results by achieving F-measures of 76.76%, 87.%, and 80.5% on the SenWave, AraCOVID19-SSD, and ArCovidVac datasets, respectively.

Social media has recently become a rich resource in mining user sentiments. In this paper, Twitter has been chosen as a platform for opinion mining in trading strategy with Mubasher products, which is a leading stock analysis software provider in the Gulf region. This experiment proposes a model for sentiment analysis of Saudi Arabic (standard and Arabian Gulf dialect) tweets to extract feedback from Mubasher products. A hybrid of natural language processing and machine learning approaches on building models are used to classify tweets according to their sentiment polarity into one of the classes positive, negative and neutral. Firstly, document's Pre-processing are explored on the dataset. Secondly, Naive Bayes and Support Vector Machines (SVMs) are applied with different feature selection schemes like TF-IDF (Term Frequency-Inverse Document Frequency) and BTO (Binary-Term Occurrence). Thirdly, the proposed model for sentiment analysis is expanded to obtain the results for N-Grams term of tokens. Finally, human has labelled the data and this may involve some mistakes in the labelling process. At this moment, neutral class with generalisation of our classification will take results to different classification accuracy.

Opinion mining is an important step towards facilitating information in health data. Several studies have demonstrated the possibility of tracking diseases using public tweets. However, most studies were applied to English language tweets. Influenza is currently one of the world's greatest infectious disease challenges. In this study, a new approach is proposed in order to detect Influenza using machine learning techniques from Arabic tweets in Arab countries. This paper is the first study of epidemic diseases based on Arabic language tweets. In this work, we have collected, labeled, filtered and analyzed the influenza-related tweets written in the Arabic language. Several classifiers were used to measure the quality and the performance of the approach, which are: Naive Bayes, Support Vector Machines, Decision Trees, and K-Nearest Neighbor. The classifiers which achieved the best accuracy results for the three experiments were: Naïve Bayes with 89.06%, and K-Nearest Neighbor with 86.43%, respectively.

In this paper, an investigation was done to identify writing style features that can be used for cross-topic and cross-genre documents in the Authorship Identification task from 2003 to 2015. Different writing style features were empirically evaluated that were previously used in single topic and single genre documents for Authorship Identification to determine whether they can be used effectively for cross-topic and crossgenre Authorship Identification using an ablation process. The dataset used was taken from the 2015 PAN CLEF Forum English collection consisting of 100 sets. Furthermore, it was investigated whether combining some of these feature sets can help improve the authorship identification task. Three different classifiers were used: Naïve Bayes, Support Vector Machine, and Random Forest. The results suggest that a combination of a lexical, syntactical, structural, and content feature set can be used effectively for cross topic and cross genre authorship identification, as it achieved an AUC result of 0.837.

Sentiment analysis mainly studies the emotional tendencies of texts from grammar, semantic rules and other aspects. The texts from social network are characterized by less words, irregular grammar, data noise and so on, which have increased the difficulty of emotion analysis. In order to improve the performance of machine learning in sentiment analysis, this study proposed the Majority Decision Algorithm to classify the emotional tendentious of the text in WeChat, combined the characteristics of five classifiers and integrated the classification results of five classifiers, eventually the text can be classified in WeChat. Firstly, this study utilized the BlueStacks to crawl the cache of WeChat Moment developed by Tencent company. Secondly, the cache was processed by Python to get the WeChat dataset. After the Chinese word segmentation, data cleaning and segmentation, the sentiment classification experiment were carried out using different classifiers. Finally, a Majority Decision Algorithm composed of five classifiers was established. It included, Naive Bayes (sklearn), Naive Bayes (SnowNLP), SVM (linear), SVM (RBF) and SGD. Then, the comparison was carried out between the performance of the algorithm and the five classifiers. Results show that the precision rates of the five classifiers are 0.8598, 0.8154, 0.8511, 0.8739 and 0.8678; the recall rates are 0.8544, 0.8482, 0.9380, 0.9226 and 0.9349; F1 scores are 0.8571, 0.8315, 0.8924, 0.8975 and 0.9001, respectively. The algorithm of the Precision rate, Recall rate and F1 score were 0.8804, 0.9349 and 0.9069, respectively, indicating that algorithm in current study significantly improved the performance, which can be effectively applied into the new text form of WeChat Moment. The study can provide theoretical reference for sentiment classification of Chinese text based on machine learning.

In this paper, we investigate the problem of recognizing entities which are targeted by text sentiment in Arabic tweets. To do so, we train a bidirectional LSTM deep neural network with conditional random fields as a classification layer on top of the network to discover the features of this specific set of entities and extract them from Arabic tweets. We’ve evaluated the network performance against a baseline method which makes use of a regular named entity recognizer and a sentiment analyzer. The deep neural network has shown a noticeable advantage in extracting sentiment target entities from Arabic tweets.

Long coronavirus disease (COVID) is a complex and multifaceted health condition with a range of severe symptoms that can last for weeks or even months after the acute phase of the illness has passed. Employing social network analysis (SNA) can rapidly provide significant health information to communities related to long COVID. This study aimed to identify the key themes, most influential users, and overall sentiments in the Twitter discourse on long COVID. Data were collected from a Twitter search with the specific keywords "long COVID" from December 1, 2022, to February 22, 2023, using NodeXL Pro. Visualizations, including network graphs and key influencers, were created using Gephi, and sentiment analysis was conducted with Azure Machine. In total, 119,185 tweets from 94325 users were related to long COVID. Top influencers include medical professionals, researchers, journalists, and public figures, with news media platforms as primary information sources; the most common hashtag was #longCOVID, indicating that it is a significant issue of concern among the Twitter community. In the sentiment analysis, most tweets were negative. The study highlights the importance of critically evaluating information shared by influential users and seeking out multiple sources of information when making health-related decisions. In addition, it emphasizes the value of examining social media conversations to understand public discourse on long COVID and suggests that future researchers could explore the role of social media in shaping public perceptions and behaviors related to health issues. Strategies for enhancing scientific journal engagement and influence in online discussions are discussed as well.

The World Health Organization (WHO) declared the outbreak of Coronavirus disease 2019 (COVID-19) a pandemic on 11 March 2020. The evolution of this pandemic has raised global health concerns, making people worry about how to protect themselves and their families. This has greatly impacted people’s sentiments. There was a dire need to investigate a large amount of social data such as tweets and others that emerged during the post-pandemic era for the assessment of people’s sentiments. As a result, this study aims at Arabic tweet-based sentiment analysis considering the COVID-19 pandemic in Saudi Arabia. The datasets have been collected in two different periods in three major regions in Saudi Arabia, which are: Riyadh, Dammam, and Jeddah. Tweets were annotated with three sentiments: positive, negative, and neutral after due pre-processing. Convolutional neural networks (CNN) and bi-directional long short memory (BiLSTM) deep learning algorithms were applied for classifying the sentiment of Arabic tweets. This experiment showed that the performance of CNN achieved 92.80% accuracy. The performance of BiLSTM was scored at 91.99% in terms of accuracy. Moreover, as an outcome of this study, an overwhelming upsurge in negative sentiments were observed in the dataset during COVID-19 compared to the negative sentiments of the dataset before COVID-19. The technique has been compared with the state-of-the-art techniques in the literature and it was observed that the proposed technique is promising in terms of various performance parameters.

PurposeThe purpose of this paper is to analyze topics as alternative features for sentiment analysis in Indonesian tweets.Design/methodology/approachGiven Indonesian tweets, the processes of sentiment analysis start by extracting features from the tweets. The features are words or topics. The authors use non-negative matrix factorization to extract the topics and apply a support vector machine to classify the tweets into its sentiment class.FindingsThe authors analyze the accuracy using the two-class and three-class sentiment analysis data sets. Both data sets are about sentiments of candidates for Indonesian presidential election. The experiments show that the standard word features give better accuracies than the topics features for the two-class sentiment analysis. Moreover, the topic features can slightly improve the accuracy of the standard word features. The topic features can also improve the accuracy of the standard word features for the three-class sentiment analysis.Originality/valueThe standard textual data representation for sentiment analysis using machine learning is bag of word and its extensions mainly created by natural language processing. This paper applies topics as novel features for the machine learning-based sentiment analysis in Indonesian tweets.

We introduce Big Data Analytics (BDA) and Sentiment Analysis (SA) to the study of international negotiations, through an application to the case of the UK-EU Brexit negotiations and the use of Twitter user sentiment. We show that SA of tweets has potential as a real-time barometer of public sentiment towards negotiating outcomes to inform government decision-making. Despite the increasing need for information on collective preferences regarding possible negotiating outcomes, negotiators have been slow to capitalise on BDA. Through SA on a corpus of 13,018,367 tweets on defined Brexit hashtags, we illustrate how SA can provide a platform for decision-makers engaged in international negotiations to grasp collective preferences. We show that BDA and SA can enhance decision-making and strategy in public policy and negotiation contexts of the magnitude of Brexit. Our findings indicate that the preferred or least preferred Brexit outcomes could have been inferred by the emotions expressed by Twitter users. We argue that BDA can be a mechanism to map the different options available to decision-makers and bring insights to and inform their decision-making. Our work, thereby, proposes SA as part of the international negotiation toolbox to remedy for the existing informational gap between decision makers and citizens’ preferred outcomes.