Polarity shift detection,elimination and ensemble: A three-stage model for document-level sentiment analysis

The polarity shift problem is a major factor that affects classification performance of machine-learning-based sentiment analysis systems. In this paper, we propose a three-stage cascade model to address the polarity shift problem in the context of document-level sentiment classification. We first split each document into a set of subsentences and build a hybrid model that employs rules and statistical methods to detect explicit and implicit polarity shifts, respectively. Secondly, we propose a polarity shift elimination method, to remove polarity shift in negations. Finally, we train base classifiers on training subsets divided by different types of polarity shifts, and use a weighted combination of the component classifiers for sentiment classification. The results on a range of experiments illustrate that our approach significantly outperforms several alternative methods for polarity shift detection and elimination.     

Self-training from labeled features for sentiment analysis

Sentiment analysis concerns about automatically identifying sentiment or opinion expressed in a given piece of text. Most prior work either use prior lexical knowledge defined as sentiment polarity of words or view the task as a text classification problem and rely on labeled corpora to train a sentiment classifier. While lexicon-based approaches do not adapt well to different domains, corpus-based approaches require expensive manual annotation effort.     

Attention-based long short-term memory network using sentiment lexicon embedding for aspect-level sentiment analysis in Korean

Although deep learning breakthroughs in NLP are based on learning distributed word representations by neural language models, these methods suffer from a classic drawback of unsupervised learning techniques. Furthermore, the performance of general-word embedding has been shown to be heavily task-dependent. To tackle this issue, recent researches have been proposed to learn the sentiment-enhanced word vectors for sentiment analysis. However, the common limitation of these approaches is that they require external sentiment lexicon sources and the construction and maintenance of these resources involve a set of complexing, time-consuming, and error-prone tasks. In this regard, this paper proposes a method of sentiment lexicon embedding that better represents sentiment word's semantic relationships than existing word embedding techniques without manually-annotated sentiment corpus. The major distinguishing factor of the proposed framework was that joint encoding morphemes and their POS tags, and training only important lexical morphemes in the embedding space. To verify the effectiveness of the proposed method, we conducted experiments comparing with two baseline models. As a result, the revised embedding approach mitigated the problem of conventional context-based word embedding method and, in turn, improved the performance of sentiment classification.     

QMOS: Query-based multi-documents opinion-oriented summarization

Sentiment analysis concerns the study of opinions expressed in a text. This paper presents the QMOS method, which employs a combination of sentiment analysis and summarization approaches. It is a lexicon-based method to query-based multi-documents summarization of opinion expressed in reviews.QMOS combines multiple sentiment dictionaries to improve word coverage limit of the individual lexicon. A major problem for a dictionary-based approach is the semantic gap between the prior polarity of a word presented by a lexicon and the word polarity in a specific context. This is due to the fact that, the polarity of a word depends on the context in which it is being used. Furthermore, the type of a sentence can also affect the performance of a sentiment analysis approach. Therefore, to tackle the aforementioned challenges, QMOS integrates multiple strategies to adjust word prior sentiment orientation while also considers the type of sentence. QMOS also employs the Semantic Sentiment Approach to determine the sentiment score of a word if it is not included in a sentiment lexicon.On the other hand, the most of the existing methods fail to distinguish the meaning of a review sentence and user's query when both of them share the similar bag-of-words; hence there is often a conflict between the extracted opinionated sentences and users’ needs. However, the summarization phase of QMOS is able to avoid extracting a review sentence whose similarity with the user's query is high but whose meaning is different. The method also employs the greedy algorithm and query expansion approach to reduce redundancy and bridge the lexical gaps for similar contexts that are expressed using different wording, respectively. Our experiment shows that the QMOS method can significantly improve the performance and make QMOS comparable to other existing methods.     

A comparative assessment of sentiment analysis and star ratings for consumer reviews

Electronic word of mouth (eWOM) is prominent and abundant in consumer domains. Both consumers and product/service providers need help in understanding and navigating the resulting information spaces, which are vast and dynamic. The general tone or polarity of reviews, blogs or tweets provides such help. In this paper, we explore the viability of automatic sentiment analysis (SA) for assessing the polarity of a product or a service review. To do so, we examine the potential of the major approaches to sentiment analysis, along with star ratings, in capturing the true sentiment of a review. We further model contextual factors (specifically, product type and review length) as two moderators affecting SA accuracy. The results of our analysis of 900 reviews suggest that different tools representing the main approaches to SA display differing levels of accuracy, yet overall, SA is very effective in detecting the underlying tone of the analyzed content, and can be used as a complement or an alternative to star ratings. The results further reveal that contextual factors such as product type and review length, play a role in affecting the ability of a technique to reflect the true sentiment of a review.     

Supervised sentiment analysis in Czech social media

This article describes in-depth research on machine learning methods for sentiment analysis of Czech social media. Whereas in English, Chinese, or Spanish this field has a long history and evaluation datasets for various domains are widely available, in the case of the Czech language no systematic research has yet been conducted. We tackle this issue and establish a common ground for further research by providing a large human-annotated Czech social media corpus. Furthermore, we evaluate state-of-the-art supervised machine learning methods for sentiment analysis. We explore different pre-processing techniques and employ various features and classifiers. We also experiment with five different feature selection algorithms and investigate the influence of named entity recognition and preprocessing on sentiment classification performance. Moreover, in addition to our newly created social media dataset, we also report results for other popular domains, such as movie and product reviews. We believe that this article will not only extend the current sentiment analysis research to another family of languages, but will also encourage competition, potentially leading to the production of high-end commercial solutions.     

Topic Driven Adaptive Network for cross-domain sentiment classification

As a hot spot these years, cross-domain sentiment classification aims to learn a reliable classifier using labeled data from a source domain and evaluate the classifier on a target domain. In this vein, most approaches utilized domain adaptation that maps data from different domains into a common feature space. To further improve the model performance, several methods targeted to mine domain-specific information were proposed. However, most of them only utilized a limited part of domain-specific information. In this study, we first develop a method of extracting domain-specific words based on the topic information derived from topic models. Then, we propose a Topic Driven Adaptive Network (TDAN) for cross-domain sentiment classification. The network consists of two sub-networks: a semantics attention network and a domain-specific word attention network, the structures of which are based on transformers. These sub-networks take different forms of input and their outputs are fused as the feature vector. Experiments validate the effectiveness of our TDAN on sentiment classification across domains. Case studies also indicate that topic models have the potential to add value to cross-domain sentiment classification by discovering interpretable and low-dimensional subspaces.     

Implicit aspect extraction in sentiment analysis: Review,taxonomy, oppportunities,and open challenges

Sentiment analysis is a text classification branch, which is defined as the process of extracting sentiment terms (i.e. feature/aspect, or opinion) and determining their opinion semantic orientation. At aspect level, aspect extraction is the core task for sentiment analysis which can either be implicit or explicit aspects. The growth of sentiment analysis has resulted in the emergence of various techniques for both explicit and implicit aspect extraction. However, majority of the research attempts targeted explicit aspect extraction, which indicates that there is a lack of research on implicit aspect extraction. This research provides a review of implicit aspect/features extraction techniques from different perspectives. The first perspective is making a comparison analysis for the techniques available for implicit term extraction with a brief summary of each technique. The second perspective is classifying and comparing the performance, datasets, language used, and shortcomings of the available techniques. In this study, over 50 articles have been reviewed, however, only 45 articles on implicit aspect extraction that span from 2005 to 2016 were analyzed and discussed. Majority of the researchers on implicit aspects extraction rely heavily on unsupervised methods in their research, which makes about 64% of the 45 articles, followed by supervised methods of about 27%, and lastly semi-supervised of 9%. In addition, 25 articles conducted the research work solely on product reviews, and 5 articles conducted their research work using product reviews jointly with other types of data, which makes product review datasets the most frequently used data type compared to other types. Furthermore, research on implicit aspect features extraction has focused on English and Chinese languages compared to other languages. Finally, this review also provides recommendations for future research directions and open problems.     

Towards sentiment and Temporal Aided Stance Detection of climate change tweets

Climate change has become one of the most significant crises of our time. Public opinion on climate change is influenced by social media platforms such as Twitter, often divided into believers and deniers. In this paper, we propose a framework to classify a tweet’s stance on climate change (denier/believer). Existing approaches to stance detection and classification of climate change tweets either have paid little attention to the characteristics of deniers’ tweets or often lack an appropriate architecture. However, the relevant literature reveals that the sentimental aspects and time perspective of climate change conversations on Twitter have a major impact on public attitudes and environmental orientation. Therefore, in our study, we focus on exploring the role of temporal orientation and sentiment analysis (auxiliary tasks) in detecting the attitude of tweets on climate change (main task). Our proposed framework STASY integrates word- and sentence-based feature encoders with the intra-task and shared-private attention frameworks to better encode the interactions between task-specific and shared features. We conducted our experiments on our novel curated climate change CLiCS dataset (2465 denier and 7235 believer tweets), two publicly available climate change datasets (ClimateICWSM-2022 and ClimateStance-2022), and two benchmark stance detection datasets (SemEval-2016 and COVID-19-Stance). Experiments show that our proposed approach improves stance detection performance (with an average improvement of 12.14% on our climate change dataset, 15.18% on ClimateICWSM-2022, 12.94% on ClimateStance-2022, 19.38% on SemEval-2016, and 35.01% on COVID-19-Stance in terms of average F1 scores) by benefiting from the auxiliary tasks compared to the baseline methods.     

News-based business sentiment and its properties as an economic index

This paper presents an approach to measuring business sentiment based on textual data. Business sentiment has been measured by traditional surveys, which are costly and time-consuming to conduct. To address the issues, we take advantage of daily newspaper articles and adopt a self-attention-based model to define a business sentiment index, named S-APIR, where outlier detection models are investigated to properly handle various genres of news articles. Moreover, we propose a simple approach to temporally analyzing how much any given event contributed to the predicted business sentiment index. To demonstrate the validity of the proposed approach, an extensive analysis is carried out on 12 years’ worth of newspaper articles. The analysis shows that the S-APIR index is strongly and positively correlated with established survey-based index (up to correlation coefficient $r=0.937$) and that the outlier detection is effective especially for a general newspaper. Also, S-APIR is compared with a variety of economic indices, revealing the properties of S-APIR that it reflects the trend of the macroeconomy as well as the economic outlook and sentiment of economic agents. Moreover, to illustrate how S-APIR could benefit economists and policymakers, several events are analyzed with respect to their impacts on business sentiment over time.     

Bi-view semi-supervised active learning for cross-lingual sentiment classification

Recently, sentiment classification has received considerable attention within the natural language processing research community. However, since most recent works regarding sentiment classification have been done in the English language, there are accordingly not enough sentiment resources in other languages. Manual construction of reliable sentiment resources is a very difficult and time-consuming task. Cross-lingual sentiment classification aims to utilize annotated sentiment resources in one language (typically English) for sentiment classification of text documents in another language. Most existing research works rely on automatic machine translation services to directly project information from one language to another. However, different term distribution between original and translated text documents and translation errors are two main problems faced in the case of using only machine translation. To overcome these problems, we propose a novel learning model based on active learning and semi-supervised co-training to incorporate unlabelled data from the target language into the learning process in a bi-view framework. This model attempts to enrich training data by adding the most confident automatically-labelled examples, as well as a few of the most informative manually-labelled examples from unlabelled data in an iterative process. Further, in this model, we consider the density of unlabelled data so as to select more representative unlabelled examples in order to avoid outlier selection in active learning. The proposed model was applied to book review datasets in three different languages. Experiments showed that our model can effectively improve the cross-lingual sentiment classification performance and reduce labelling efforts in comparison with some baseline methods.     

Hierarchical template transformer for fine-grained sentiment controllable generation

Existing methods for text generation usually fed the overall sentiment polarity of a product as an input into the seq2seq model to generate a relatively fluent review. However, these methods cannot express more fine-grained sentiment polarity. Although some studies attempt to generate aspect-level sentiment controllable reviews, the personalized attribute of reviews would be ignored. In this paper, a hierarchical template-transformer model is proposed for personalized fine-grained sentiment controllable generation, which aims to generate aspect-level sentiment controllable reviews with personalized information. The hierarchical structure can effectively learn sentiment information and lexical information separately. The template transformer uses a part of speech (POS) template to guide the generation process and generate a smoother review. To verify our model, we used the existing model to obtain a corpus named FSCG-80 from Yelp, which contains 800K samples and conducted a series of experiments on this corpus. Experimental results show that our model can achieve up to 89.93% aspect-sentiment control accuracy and generate more fluent reviews.     

Going beyond fact-checking to fight health misinformation: A multi-level analysis of the Twitter response to health news stories

Health misinformation has become an unfortunate truism of social media platforms, where lies could spread faster than truth. Despite considerable work devoted to suppressing fake news, health misinformation, including low-quality health news, persists and even increases in recent years. One promising approach to fighting bad information is studying the temporal and sentiment effects of health news stories and how they are discussed and disseminated on social media platforms like Twitter. As part of the effort of searching for innovative ways to fight health misinformation, this study analyzes a dataset of more than 1600 objectively and independently reviewed health news stories published over a 10-year span and nearly 50,000 Twitter posts responding to them. Specifically, it examines the source credibility of health news circulated on Twitter and the temporal, sentiment features of the tweets containing or responding to the health news reports. The results show that health news stories that are rated low by experts are discussed more, persist longer, and produce stronger sentiments than highly rated ones in the tweetosphere. However, the highly rated stories retained a fresh interest in the form of new tweets for a longer period. An in-depth understanding of the characteristics of health news distribution and discussion is the first step toward mitigating the surge of health misinformation. The findings provide insights into understanding the mechanism of health information dissemination on social media and practical implications to fight and mitigate health misinformation on digital media platforms.     

Social media mining for product planning: A product opportunity mining approach based on topic modeling and sentiment analysis

Social media data have recently attracted considerable attention as an emerging voice of the customer as it has rapidly become a channel for exchanging and storing customer-generated, large-scale, and unregulated voices about products. Although product planning studies using social media data have used systematic methods for product planning, their methods have limitations, such as the difficulty of identifying latent product features due to the use of only term-level analysis and insufficient consideration of opportunity potential analysis of the identified features. Therefore, an opportunity mining approach is proposed in this study to identify product opportunities based on topic modeling and sentiment analysis of social media data. For a multifunctional product, this approach can identify latent product topics discussed by product customers in social media using topic modeling, thereby quantifying the importance of each product topic. Next, the satisfaction level of each product topic is evaluated using sentiment analysis. Finally, the opportunity value and improvement direction of each product topic from a customer-centered view are identified by an opportunity algorithm based on product topics’ importance and satisfaction. We expect that our approach for product planning will contribute to the systematic identification of product opportunities from large-scale customer-generated social media data and will be used as a real-time monitoring tool for changing customer needs analysis in rapidly evolving product environments.     

How CEOs use Twitter: A comparative analysis of Global and Latin American companies

Social networks are becoming a key communication tool for organizations, but also for top managers like CEOs. Among the different available platforms, Twitter is one of the greatest and it is considered one of the most suitable to share information and engage in dialogue with stakeholders. In this way, this paper analyzes the presence of CEOs on the most active social network sites, and assess the activity and interaction of these top managers on Twitter. CEOs from Global and Latin American companies were selected, to compare their performance. The results of the study show that the presence of CEOs in social networks is very low, and the majority of those that are present on them are not adequately using their Twitter accounts. Although the general presence and performance on are low, LatAm CEOs have a better presence on social networks and they are more active on Twitter, but Global CEOs have better interaction results on their accounts. So, this area of strategic communication should be improved by communication practitioners, since the CEO communication is nowadays a key communication issue for any organization.     

Cognitive-inspired domain adaptation of sentiment lexicons

Sentiment lexicons are essential tools for polarity classification and opinion mining. In contrast to machine learning methods that only leverage text features or raw text for sentiment analysis, methods that use sentiment lexicons embrace higher interpretability. Although a number of domain-specific sentiment lexicons are made available, it is impractical to build an ex ante lexicon that fully reflects the characteristics of the language usage in endless domains. In this article, we propose a novel approach to simultaneously train a vanilla sentiment classifier and adapt word polarities to the target domain. Specifically, we sequentially track the wrongly predicted sentences and use them as the supervision instead of addressing the gold standard as a whole to emulate the life-long cognitive process of lexicon learning. An exploration-exploitation mechanism is designed to trade off between searching for new sentiment words and updating the polarity score of one word. Experimental results on several popular datasets show that our approach significantly improves the sentiment classification performance for a variety of domains by means of improving the quality of sentiment lexicons. Case-studies also illustrate how polarity scores of the same words are discovered for different domains.     

Incorporating sentiment into tag-based user profiles and resource profiles for personalized search in folksonomy

In recent years, there has been a rapid growth of user-generated data in collaborative tagging (a.k.a. folksonomy-based) systems due to the prevailing of Web 2.0 communities. To effectively assist users to find their desired resources, it is critical to understand user behaviors and preferences. Tag-based profile techniques, which model users and resources by a vector of relevant tags, are widely employed in folksonomy-based systems. This is mainly because that personalized search and recommendations can be facilitated by measuring relevance between user profiles and resource profiles. However, conventional measurements neglect the sentiment aspect of user-generated tags. In fact, tags can be very emotional and subjective, as users usually express their perceptions and feelings about the resources by tags. Therefore, it is necessary to take sentiment relevance into account into measurements. In this paper, we present a novel generic framework SenticRank to incorporate various sentiment information to various sentiment-based information for personalized search by user profiles and resource profiles. In this framework, content-based sentiment ranking and collaborative sentiment ranking methods are proposed to obtain sentiment-based personalized ranking. To the best of our knowledge, this is the first work of integrating sentiment information to address the problem of the personalized tag-based search in collaborative tagging systems. Moreover, we compare the proposed sentiment-based personalized search with baselines in the experiments, the results of which have verified the effectiveness of the proposed framework. In addition, we study the influences by popular sentiment dictionaries, and SenticNet is the most prominent knowledge base to boost the performance of personalized search in folksonomy.     

Investigating the COVID-19 vaccine discussions on Twitter through a multilayer network-based approach

Modeling discussions on social networks is a challenging task, especially if we consider sensitive topics, such as politics or healthcare. However, the knowledge hidden in these debates helps to investigate trends and opinions and to identify the cohesion of users when they deal with a specific topic. To this end, we propose a general multilayer network approach to investigate discussions on a social network. In order to prove the validity of our model, we apply it on a Twitter dataset containing tweets concerning opinions on COVID-19 vaccines. We extract a set of relevant hashtags (i.e., gold-standard hashtags) for each line of thought (i.e., pro-vaxxer, neutral, and anti-vaxxer). Then, thanks to our multilayer network model, we figure out that the anti-vaxxers tend to have ego networks denser (+14.39%) and more cohesive (+64.2%) than the ones of pro-vaxxer, which leads to a higher number of interactions among anti-vaxxers than pro-vaxxers (+393.89%). Finally, we report a comparison between our approach and one based on single networks analysis. We prove the effectiveness of our model to extract influencers having ego networks with more nodes (+40.46%), edges (+39.36%), and interactions with their neighbors (+28.56%) with respect to the other approach. As a result, these influential users are much more important to analyze and can provide more valuable information.