Top Research Papers on Sentiment Analysis

A first of its kind overview of key sentiment analysis techniques and eXplainable artificial intelligence (XAI) methodologies that are currently in use and a comprehensive review of sentiment analysis explainability.

This work proposes two strategies for defining sentence difficulty, which are binary and considers sentences as difficult when the classifiers are unable to correctly assign the sentiment polarity, and uses a six-level difficulty scale.

A new Multimodal Representation Learning (MRL) method for Multimodal Sentiment Analysis (MSA), which facilitates the adaptive interaction between modalities through Cooperative Sentiment Agents, named Co-SA, which excels at discovering diverse cross-modal features.

This study classifies YouTube comments, investigates sentiment analysis techniques that can be used on them, and offers insightful information useful for data mining and sentiment analysis research.

This work introduces the first large-scale human-annotated Twitter sentiment dataset for Nigeria—Hausa, Igbo, Nigerian-Pidgin, and Yorùbá—consisting of around 30,000 annotated tweets per language, including a significant fraction of code-mixed tweets.

The results showed that neutral sentiment was the highest result with a percentage of 48% followed by negative sentiment of 23%, positive sentiment 22%, high positive sentiment 4%, and high negative sentiment 3%.

This Element provides a basic introduction to sentiment analysis, aimed at helping students and professionals in corpus linguistics to understand what sentiment analysis is, how it is conducted, and where it can be applied.

An exhaustive survey of the latest improvements around here, and a large number of recently proposed calculations and applications are investigated.

An ABSA method that integrates explicit sentiment augmentations (ABSA-ESA) to add more sentiment clues and the results show that ABSA-ESA outperforms the SOTA baselines on implicit and explicit sentiment accuracy.

This review provides an overview of the definition, background, and development of multimodal sentiment analysis, and covers recent datasets and advanced models, emphasizing the challenges and future prospects of this technology.

The trend of text-based emotion detection has shifted from the early keyword-based comparisons to machine learning and deep learning algorithms that provide more flexibility to the task and better performance.

This study proposes text classification with the use of bidirectional encoder representations from transformers (BERT) for natural language processing with other variants to improve accuracy rate, precision rate, recall rate, and F1-score.

An ethics sheet encompassing critical inquiries to guide practitioners in ensuring equitable utilization of sentiment analysis is proposed to underscore the significance of adopting an interdisciplinary approach to defining sentiment in SA.

This research focuses on developing a social media sentiment analysis framework, incorporating a custom-built emotion thesaurus to enhance the precision of sentiment analysis, and investigates the efficacy of various deep learning algorithms, under different parameter calibrations, for sentiment extraction from social media.

This paper retrieves and analyzes relevant literature on text sentiment analysis from the past decade, performing a thematic analysis to summarize and categorize the mainstream methods used in sentiment analysis, and discusses their strengths and weaknesses.

An overview of how and where Sentiment analysis is used in the agrarian sector and which methods are most commonly used is provided.

A clearer scope for FSA studies is defined and the FSA-investor sentiment-market sentiment relationship is conceptualized, which defines a clearer scope for FSA studies and conceptualizes the FSA-investor sentiment-market sentiment relationship.

The authors propose sentiment analysis models based on naive Bayes, logistic regression, and support vector machine algorithms with the objective of achieving more effective sentiment analysis, utilizing a publicly available labeled dataset from Kaggle.

The paper analyzes the accuracy of the Azure Sentiment Analysis service for five languages, namely Romanian, French, Italian, Portuguese, and Spanish, and suggests that texts of moderate length are easier to classify.

This research proposes a multimodal framework that integrates visual and textual features to predict the GIF sentiment and incorporates attributes including face emotion detection and OCR generated captions to capture the semantic aspects of the GIF.

A novel model called Continuous Sentiment Contextualized Vectors (CSCV) is introduced to address the problem of sentiment analysis, which uses Continuous Bag-of-Words model to deal with the context and sentiment lexicons to identify sentiment.

People can be able to view the news and know whether the news is positive, neutral or negative by knowing sentiment analysis of the particular news through one of the sentiment analysis techniques called lexicon sentiment and using Deep learning algorithms.

Twitter is an enormously popular microblog on which clients may voice their opinions and opinion investigation of Twitter data is a field that has been given much attention over the last decade and involves dissecting “tweets” (comments) and the content of these expressions.

This work proposes the Sentiment Interaction Distillation (SID) Network, which utilizes object sentimental interaction to guide feature learning and proposes a knowledge distillation framework to utilize interaction information guiding global context feature learning.

This work explores an approach utilizing state-of-the-art machine translation technology and performs sentiment analysis on the English translation of a foreign language text and indicates that entity sentiment scores obtained by the method are statistically significantly correlated across nine languages of news sources and five languages of a parallel corpus.

A platform that collect comments from the user and creates classified and structured overviews of such comments and facilitates access to its data is introduced.

There is a tendency for positive perceptions of Halal Certification in the opinion of Twitter users, with positive sentiment of 41.8%, neutral sentiment of 30.8% and negative sentiment of 27.4%.

This paper obtained the sentiment information of words under optimal sentiment concept from the multi-semantics sentiment intensity lexicon which was constructed in this paper to achieve accurate embedding of sentiment information and provide more accurate semantics and sentiment representation for words.

Sentiment analysis or opinion mining is the computational study of people’s opinions, appraisals, and emotions toward entities, events and their attributes in order to gather consumer opinions about its products and those of its competitors.

Experimental comparisons show that automated idiom enrichment and annotation are very beneficial for the performance of the sentiment classifier and the expanded annotated lexicon will be made available to the general public.

This study aims to evaluate sentiments related to profit management practices using secondary data obtained from articles that have been published on the Scopus database. A qualitative approach is applied in this research methodology, with the utilization of SentiStrength software to classify sentiments. The results of the analysis show that positive sentiment dominates at 40% towards profit management, followed by neutral sentiment at 37%, and negative sentiment at 23%. Some of the positive aspects include improving economic welfare, investment and development, human resource development, and...

Positive correlations between the number of tweets and the daily prices, and between the prices of different crypto coins, are found.

The use of textual analysis and sentiment analysis in accounting is reviewed and a review of the previous literature on the use of these techniques in finance and accounting is performed and the main techniques of sentiment analysis are described.

This work examines the behaviour of an aspect-based sentiment classifier built bytuning the 002 BERT BASE model on the SemEval 2016 En-003 glish dataset and uses a gradient-based method to identify the most salient words.

The purpose of this study is to analyze the sentiment resulting from the opinion of researchers through published articles related to cash waqf. The data analyzed consisted of 107 research articles indexed by Scopus in 2022. The method used is a qualitative approach with descriptive statistics from literature studies on the theme of cash waqf. The results showed that neutral sentiment was the highest with a percentage of 44.9%, followed by positive sentiment of 31.8%, negative sentiment of 23.4%. The diversity of sentiment results is obtained because of the pros and cons of cash waqf.

From defining the sentiment analysis to algorithms for sentiment analysis and from the first step of sentiment analysis to evaluating the predictions of sentiment classifiers, additional feature extractions to boost performance are discussed with practical results.

This paper is an attempt to perform the sentiment analysis on the Urdu language by extracting the tweets (Nastaleeq and Roman Urdu both) from Twitter using Tweepy API and the tool opted for the purpose is WEKA.

Product Review Sentiment Analysis is a fundamental aspect of contemporary business strategies, utilizing Machine Learning techniques to glean insights from vast textual datasets to glean insights from vast textual datasets.

The experimental results show that the proposed KWS based system significantly outperforms the traditional ASR architecture in detecting sentiment for challenging practical tasks.

The BERT transformer method with an Accuracy: 0.973 is recommended for sentiment analysis and the Naive Bayes (NB), and K-Nearest Neighbor (KNN) models are used.

Purpose – This study aims to determine the sentiment towards women's entrepreneurship development in the last 32 years, from 1987 to 2019Methodology – The method used is descriptive statistical analysis combined with meta-analysis and sentiment from secondary data in metadata from 98 Scopus indexed publications, then processed using Microsoft Excel 2019 and SentiStrength software.Findings – The sentiment analysis results show that the expert opinions on the development of women's entrepreneurship are diverse, with positive sentiment at 30.6%, negative sentiment at 30.6%, and neutral sentiment ...

The results show that people don't have positive sentiments towards this coronavirus, COVID 19.

This article study the different uses of sentiment analysis in the detection of fake news, with a discussion of the most relevant elements and shortcomings, and the requirements that should be met in the near future, such as multilingualism, explainability, mitigation of biases, or treatment of multimedia elements.

This work proposes a model that refines pre-trained word embeddings with context information and leverages the sentiment scores of sentences obtained from a lexicon-based method to further improve performance of domain-specific sentiment analysis tasks.

A new sentiment-enhanced word embedding (S-EWE) method to improve the effectiveness of sentence-level sentiment classification and takes full advantage of the mapping relationship between word embeddings and their corresponding sentiment orientations.

A thorough review of the most recent developments in Twitter Sentiment Analysis (TSA) is reviewed, and a wide range of newly proposed algorithms and applications are explored.

The proposed method can provide an increase of 3.5 times lexicon number as well as improve the accuracy of 80.9% for online review and 95.7% for Twitter data, and they are better than other published and available Indonesian sentiment lexicons.

SentiWSP is proposed, a novel Sentiment-aware pre-trained language model with combined Word-level and Sentence-level Pre-training tasks that achieves new state-of-the-art performance on various sentence- level and aspect-level sentiment classification benchmarks.

This work proposes training the visual sentiment analysis model by distilling the result of text classification that takes the pseudo text derived from patch images of object instances, which is lower than the original SentiBank in the Twitter dataset, but achieved higher accuracy in the EmotionROI dataset.

The study addresses the unique challenges of sentiment analysis in the context of online product evaluations, including polarity changes, sarcasm, and domain-specific sentiment expressions, which often pose significant obstacles to precise sentiment classification.