Article
Computer Science, Information Systems
Belgacem Brahimi, Mohamed Touahria, Abdelkamel Tari
Summary: This paper proposes methods to extract valuable opinions from online movie reviews using n-gram and skip-n-gram models, subjective words, and feature reduction techniques to enhance sentiment analysis in Arabic. Experimental results demonstrate the effectiveness of these methods in improving sentiment classification results.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2021)
Article
Computer Science, Artificial Intelligence
Baris Ozyurt, M. Ali Akcayol
Summary: With the widespread use of social networks and other platforms, the volume of user-generated textual data is growing rapidly, making sentiment analysis and opinion mining in user reviews more and more important. To tackle issues like data sparsity and lack of co-occurrence patterns, studies have proposed methods like SS-LDA to adapt LDA for short texts. Experimental results indicate that SS-LDA performs competitively in extracting product aspects.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Review
Computer Science, Software Engineering
Bin Lin, Nathan Cassee, Alexander Serebrenik, Gabriele Bavota, Nicole Novielli, Michele Lanza
Summary: Opinion mining, also known as sentiment analysis, has gained attention in software engineering research for identifying developer emotions and extracting user criticisms in mobile apps. Through a systematic literature review of 185 papers, we provide well-defined categories of opinion mining-related software development activities, available opinion mining approaches, datasets for evaluation and tool customization, and concerns or limitations for researchers to consider when applying or customizing opinion mining techniques. Our study serves as a reference for selecting suitable opinion mining tools and provides critical insights for the further development of this technique in software engineering.
ACM TRANSACTIONS ON SOFTWARE ENGINEERING AND METHODOLOGY
(2022)
Article
Computer Science, Artificial Intelligence
Li Yang, Jieming Wang, Jin-Cheon Na, Jianfei Yu
Summary: Multimodal entity-category-sentiment triple extraction (MECSTE) is an emerging task in sentiment analysis, aiming to simultaneously extract entities, fine-grained entity categories, and sentiment polarities from sentences and images. Previous studies have overlooked the interconnection among subtasks and failed to provide sufficient information for disambiguating entities. This study proposes a generative multimodal approach and demonstrates its superiority through experiments on annotated Twitter datasets.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Giuseppe D'Aniello, Matteo Gaeta, Ilaria La Rocca
Summary: This article presents an overview of techniques and approaches for aspect-based sentiment analysis (ABSA) and highlights the main issues in this field. The KnowMIS-ABSA model is proposed, which emphasizes that sentiment, affect, emotion, and opinion are different concepts and should be measured using different tools and metrics.
ARTIFICIAL INTELLIGENCE REVIEW
(2022)
Article
Computer Science, Artificial Intelligence
Jianfei Yu, Kai Chen, Rui Xia
Summary: The study proposes a general Hierarchical Interactive Multimodal Transformer (HIMT) model to address the shortcomings in aspect-based multimodal sentiment analysis (ABMSA) and achieves significant improvements in experimental results.
IEEE TRANSACTIONS ON AFFECTIVE COMPUTING
(2023)
Article
Computer Science, Artificial Intelligence
Fang Chen, Zhongliang Yang, Yongfeng Huang
Summary: This paper proposes a novel multi-task learning framework for end-to-end aspect sentiment triplet extraction (ASTE). By decomposing ASTE into target tagging, opinion tagging, and sentiment tagging subtasks, and utilizing specific tagging schemes, our framework achieves better performance in extracting overlapping triplets and identifying long-range correspondences.
Article
Computer Science, Artificial Intelligence
Eman M. Aboelela, Walaa Gad, Rasha Ismail
Summary: In recent years, many users prefer online shopping, allowing customers to submit comments and feedback on shopping websites. Opinion mining and sentiment analysis are used to assist buyers and sellers in making purchase decisions. A semantic-based aspect level opinion mining (SALOM) model is proposed to consider negation words and other types of product aspects, with promising experimental results.
PEERJ COMPUTER SCIENCE
(2021)
Article
Computer Science, Artificial Intelligence
Zoltan Geler, Milos Savic, Brankica Bratic, Vladimir Kurbalija, Mirjana Ivanovic, Weihui Dai
Summary: This research focuses on ML-based sentiment analysis of food services reviews, comparing regression models to predict customer satisfaction. Keywords extracted from customer reviews have potential for predicting satisfaction in food taste, service, and environment aspects.
CONNECTION SCIENCE
(2021)
Article
Computer Science, Information Systems
Mohd Suhairi Md Suhaimin, Mohd Hanafi Ahmad Hijazi, Ervin Gubin Moung, Puteri Nor Ellyza Nohuddin, Stephanie Chua, Frans Coenen
Summary: The importance of social media sentiment analysis and opinion mining for public security has grown over the years. This paper presents a survey of the current state-of-the-art in this field, aiming to understand progress, identify research gaps, and propose future directions. However, there is currently a lack of systematic surveys describing the trends and latest developments in this domain.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2023)
Article
Mathematics
Azra Shamim, Muhammad Ahsan Qureshi, Farhana Jabeen, Misbah Liaqat, Muhammad Bilal, Yalew Zelalem Jembre, Muhammad Attique
Summary: This paper proposes an opinion mining system that ranks reviews and features based on novel ranking schemes and innovative visualization methods to empower users to spot imperative product features from enormous reviews, improving the decision-making process.
Article
Computer Science, Information Systems
Li Yang, Jin-Cheon Na, Jianfei Yu
Summary: In this paper, we propose a multi-task learning framework called CMMT for End-to-End Multimodal Aspect-Based Sentiment Analysis. Experimental results demonstrate that CMMT consistently outperforms the state-of-the-art approach JML and achieves superior performance in aspect extraction and sentiment classification compared to other systems.
INFORMATION PROCESSING & MANAGEMENT
(2022)
Article
Computer Science, Artificial Intelligence
Marouane Birjali, Mohammed Kasri, Abderrahim Beni-Hssane
Summary: Sentiment analysis, also known as Opinion Mining, is the task of extracting and analyzing people's opinions and emotions towards different entities. It is a powerful tool used by businesses, governments, and researchers to gain insights and make better decisions. This paper provides a comprehensive study of sentiment analysis methods, challenges, and trends for researchers in the field.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Alireza Ghorbanali, Mohammad Karim Sohrabi
Summary: "Sentiment analysis plays a vital role in natural language processing with wide-ranging applications. The rise of social media and its associated tools and technologies has led to the sharing of multimodal content and opinions in various media forms, including text, images, videos, audio, and emojis. Compared to single-modal data, multimodal data contain more valuable information for understanding users' real sentiments. Deep learning-based approaches have emerged to address the challenges of multimodal sentiment analysis, such as incomplete data, heterogeneity of modalities, fusion methods, and interactions between modals. This paper provides a comprehensive survey of sentiment analysis approaches, challenges, applications, and trends, with a particular focus on deep learning-based multimodal sentiment analysis methods."
ARTIFICIAL INTELLIGENCE REVIEW
(2023)
Article
Computer Science, Information Systems
Hiren Kumar Thakkar, Prasan Kumar Sahoo, Pranab Mohanty
Summary: This paper introduces a novel method for domain feature retrieval in text summarization, formulating the problem as a clustering problem and utilizing three newly conceived empirical observations. Two algorithms are designed to identify domain features, with experimental results demonstrating the robustness of the method in domain feature retrieval and summarization.
INFORMATION PROCESSING & MANAGEMENT
(2021)