Text Sentiment Research Articles

Deep Learning Short Text Sentiment Analysis Based on Improved Particle Swarm Optimization

Manually tuning the hyperparameters of a deep learning model is not only a time-consuming and labor-intensive process, but it can also easily lead to issues like overfitting or underfitting, hindering the model’s full convergence. To address this challenge, we present a BiLSTM-TCSA model (BiLSTM combine TextCNN and Self-Attention) for deep learning-based sentiment analysis of short texts, utilizing an improved particle swarm optimization (IPSO). This approach mimics the global random search behavior observed in bird foraging, allowing for adaptive optimization of model hyperparameters. In this methodology, an initial step involves employing a Generative Adversarial Network (GAN) mechanism to generate a substantial corpus of perturbed text, augmenting the model’s resilience to disturbances. Subsequently, global semantic insights are extracted through Bidirectional Long Short Term Memory networks (BiLSTM) processing. Leveraging Convolutional Neural Networks for Text (TextCNN) with diverse convolution kernel sizes enables the extraction of localized features, which are then concatenated to construct multi-scale feature vectors. Concluding the process, feature vector refinement and the classification task are accomplished through the integration of Self-Attention and Softmax layers. Empirical results underscore the effectiveness of the proposed approach in sentiment analysis tasks involving succinct texts containing limited information. Across four distinct datasets, our method attains impressive accuracy rates of 91.38%, 91.74%, 85.49%, and 94.59%, respectively. This performance constitutes a notable advancement when compared against conventional deep learning models and baseline approaches.

Peer-to-Peer Accommodation User Experience: Evidence From Sri Lanka

With the emergence of the sharing economy, Airbnb has become the poster child for the digital peer-to-peer accommodation market in the hospitality industry. The study explores the attributes influencing Airbnb user experiences in Sri Lanka by analyzing online guest reviews. The study uses an exploratory research design to examine a large data set of Airbnb user reviews utilizing text mining and sentiment analysis. According to the analysis, five key themes emerged from the review comments: “recommend,”“host,”“experience,”“room,” and “location.” Furthermore, the relationship between the host and the guest greatly impacts the Airbnb user experience in Sri Lanka. Most Airbnb user reviews are positive and primarily focus on the host, while negative ones typically focus on the property’s indoor environmental quality. The study contributes to the literature on the sharing economy by offering a structured and extensive analysis of actual online reviews from Airbnb customers to understand their experiences better while providing valuable insights for various stakeholders in the industry.

Big data sentiment analysis of business environment public perception based on LTP text classification ——Take Heilongjiang province as an example

At present, the research of business environment is limited to conducting surveys on specific groups or measuring data from official databases. The assessment of the business environment largely depends on public perception. Aiming to explore the public perception of business environment, this paper organically combines the big data text mining and sentiment analysis (SA). The results show that the combination of big data text mining and SA can reflect the theme characteristics, reduce the bias of sentiment and text analysis, and clearly show the public perception of the business environment. The empirical study found that the public perception of business environment depends on not only the four dimensions of business environment, but also the influence of public opinion that cannot be ignored. The public's low recognition of the business environment in Heilongjiang Province mainly includes backward economic development, serious brain drain, low government efficiency, imperfect policies, administrative law enforcement, regional climate and urban construction. In order to solve these problems, it is necessary to improve the high-standard market system to promote economic development, enhance the efficiency of government services, improve government policies, effectively enhance law enforcement, strengthen infrastructure construction and promote cultural innovation.

Social Image–Text Sentiment Classification With Cross-Modal Consistency and Knowledge Distillation

Social media sentiment analysis, which aims to evaluate the attitudes of online users based on their posts, has attracted significant research attention due to its successful application in the field of social media monitoring. It is a beneficial way to utilize multimodal information uploaded by users in order to improve sentiment classification ability. However, existing multimodal fusion-based approaches continue to face difficulties due to the issues of between-modality semantic inconsistency and missing modality. To address these issues, we propose a cross-modal consistency modeling-based knowledge distillation framework for image-text sentiment classification of social media data. Specifically, we design a hybrid curriculum learning strategy to measure the semantic consistency of multimodal data, then gradually train all image-text pairs from easy to hard, which can effectively handle the massive amounts of noise caused by inconsistencies between image and text data on social media. Moreover, in order to alleviate the problem of missing images in unimodal posts, we propose a privileged feature distillation method, in which the teacher model additionally considers images as privileged features, to transfer the visual knowledge to the student model, thereby enhancing the accuracy for text sentiment classification. Extensive experiments conducted over three real-world social media datasets demonstrate the effectiveness and superiority of the proposed multimodal sentiment analysis model.

Data augmentation for sentiment classification with semantic preservation and diversity

Data augmentation is a commonly-used technique to avoid over-fitting in deep learning. However, the mechanism behind effective data augmentation methods is unclear. To address this issue, we explore and identify two critical factors: semantic preservation and diversity to assess the quality of data augmentation in natural language processing. Our study focus on text sentiment classification and examines these two factors on two commonly-used data augmentation methods: synonym replacement and random deletion. Based on the discovery, we propose two new augmentation methods: TF-IDF word dropout and adaptive synonym replacement. Experimental results demonstrate that these two new data augmentation methods are effective. Moreover, with further experiments, we summarize three strategies for improving data augmentation methods in sentiment classification task. These strategies are employing online augmentation, introducing word importance into word sampling process, and filtering augmented data based on the current model state. We hope that our study will inspire some new perspectives on the underlying principles of data augmentation’s effectiveness and contribute to a systematic study of data augmentation methods in future.

Can textual sentiment partially explain differences in the prices of dual-listed stocks?

This study offers a robust examination of the impact of textual sentiment on price deviations for dual-listed stocks. We analysed four (three) pairs of twin stocks to establish the association between news (Twitter) sentiment differentials and theoretical price ratios. The results reveal that: Anglo-American and Anglo-Australian twin stocks exhibit a negative relationship between news (Twitter) sentiment and deviations from theoretical price parity, while the Anglo-South African twin stocks display a positive relationship using news sentiment differentials. We shed light on the role of online sentiment in understanding price deviations and offer valuable insights for investors and analysts.

Enhancing Deep Learning-Based Sentiment Analysis Using Static and Contextual Language Models

Sentiment Analysis (SA) is an essential task of Natural Language Processing and is used in various fields such as marketing, brand reputation control, and social media monitoring. The various scores generated by users in product reviews are essential feedback sources for businesses to discover their products' positive or negative aspects. However, it takes work for businesses facing a large user population to accurately assess the consistency of the scores. Recently, automated methodologies based on Deep Learning (DL), which utilize static and especially pre-trained contextual language models, have shown successful performances in SA tasks. To address the issues mentioned above, this paper proposes Multi-layer Convolutional Neural Network-based SA approaches using Static Language Models (SLMs) such as Word2Vec and GloVe and Contextual Language Models (CLMs) such as ELMo and BERT that can evaluate product reviews with ratings. Focusing on improving model inputs by using sentence representations that can store richer features, this study applied SLMs and CLMs to the inputs of DL models and evaluated their impact on SA performance. To test the performance of the proposed approaches, experimental studies were conducted on the Amazon dataset, which is publicly available and considered a benchmark dataset by most researchers. According to the results of the experimental studies, the highest classification performance was obtained by applying the BERT CLM with 82% test and 84% training accuracy scores. The proposed approaches can be applied to various domains' SA tasks and provide insightful decision-making information.

A computational approach towards food-wine recommendations

Food-wine pairing is an essential study in the culinary world and requires extensive research of the underlying food and wine pairing principles. To understand the principles of pairing, we must understand the characteristics of food, and wine, the interaction between them, and certain classic food-wine pairing norms that are religiously adhered to. To take this understanding to a broader class of masses, there is a need for a recommender system that considers all the trivial and non-trivial details of food-wine pairing and the preferences of individual users to provide the perfect pairing. To apply these pairing principles in the context of recommender systems, we require abstract features such as flavor, aroma, major ingredients, type of meat, and many others. To achieve this, our recommender system relies on text mining techniques and sentiment analysis at its core to extract the relevant information from the text in the dataset. Such extensive feature engineering has not been undertaken before in the context of generating food and wine recommendations. We have created a recommender system that embodies the age-old essence of wine paring, considering not only the food wine characteristics but also the user’s preference of pairing (congruent or contrast).

On the prediction of stock price crash risk using textual sentiment of management statement

PurposeSince stock return and volatility matters to investors, this study proposes to incorporate the textual sentiment of annual reports in stock price crash risk prediction.Design/methodology/approachSpecific sentences gathered from management discussions and their subsequent analyses are tokenized and transformed into numeric vectors using textual mining techniques, and then the Naïve Bayes method is applied to score the sentiment, which is used as an input variable for crash risk prediction. The results are compared between a collection of predictive models, including linear regression (LR) and machine learning techniques.FindingsThe experimental results find that those predictive models that incorporate textual sentiment significantly outperform the baseline models with only accounting and market variables included. These conclusions hold when crash risk is proxied by either the negative skewness of the return distribution or down-to-up volatility (DUVOL).Research limitations/implicationsIt should be noted that the authors' study focuses on examining the predictive power of textual sentiment in crash risk prediction, while other dimensions of textual features such as readability and thematic contents are not considered. More analysis is needed to explore the predictive power of textual features from various dimensions, with the most recent sample data included in future studies.Originality/valueThe authors' study provides implications for the information value of textual data in financial analysis and risk management. It suggests that the soft information contained within annual reports may prove informative in crash risk prediction, and the incorporation of textual sentiment provides an incremental improvement in overall predictive performance.

Analysis of the Sentiment in the Evaluation Texts of University Students by Means of the Concept of Flexible Management

With the development of information technology (IT) and the popularity of the Internet, it is easier to obtain college students’ teaching evaluation text data. However, it is still challenging to deeply understand and effectively analyze these texts. Based on the flexible management concept in universities, this study aimed to understand and analyze the teaching evaluation texts of college students more accurately, thereby improving the teaching quality. The implicit features of those texts were recognized first in this study, which aimed to supplement and improve the implicit features neglected by existing sentiment analysis methods. A sentiment analysis method based on Bidirectional Encoder Representations from Transformers (BERT) was adopted to explore the deep semantic information of texts. It was done by using deep learning technology and to improve the accuracy of sentiment analysis. This in turn provides more valuable reference information for university teaching management and further promotes the practical application of the flexible management concept in university teaching management.

The value of communication during pandemics

Using natural language processing (NLP) techniques to analyze the leading newspapers in China, we develop a novel measure, a pandemic intensity index. We show that firms with more investor communication have significantly higher market value during the multiple pandemic outbreaks over the period 2013 to 2021. Investor communication works primarily through alleviating information asymmetry between firms and investors and raising a firm's profile with market participants during pandemics. Our results provide support for the hypothesis that investor communication acts as a consoler, which helps firms prevail under unexpected conditions. Positive textual sentiment enhances the communication effect. We conclude that firms should enhance communication with investors and reassure them during turbulent times, especially in response to uncertainty.

Understanding sovereign credit ratings: Text-based evidence from the credit rating reports

We apply a novel approach to identifying the qualitative judgment of the rating committee in sovereign credit ratings by extending the traditional regression with new measures – sentiment and subjectivity scores – obtained by textual sentiment analysis methods. Using an ordered logit with random effects for 98 countries from 1995 to 2018, we find evidence that the subjectivity score provides additional information not captured by previously identified determinants of sovereign credit ratings, even after controlling for political risk, institutional strength, and potential bias. The results from the bivariate and multivariate analysis confirm differences in textual sentiment and subjectivity between emerging markets and advanced economies, as well as before and after the 2008 global financial crisis.

Text sentiment analysis using deep convolutional networks

The ability to process language is a difficult task. A computer system that can extract meaningful information from raw text data is considered to be equipped with intelligent behaviour. Sentiment analysis is a form of information extraction of growing research and commercial interest, especially because of the exponential increase of data in recent years. This paper presents a machine learning approach built upon Neural Networks that equips a computer system with the ability to learn how to detect sentiment in a given text and to correctly classify previously unencountered text with respect to predefined sentiment categories. The developed approach does not rely on any hard-coded parameters. Rather, a model is learned through a training procedure performed on raw data alone with the objective of searching to identify patterns and discovering feature predictors that allow the algorithm to classify text into positive, negative, or neutral sentiments, while additionally focusing on nuances such as humorous or sarcastic text. Experimental validation shows promising results. The learning algorithm is fed with high-volume textual data extracted from the Twitter social media platform. The trained model is then tested on a separate dataset, previously unknown by the model. The classification algorithm is able to predict the sentiment associated with a given text with high accuracy.

Sentiment Analysis using a CNN-BiLSTM Deep Model Based on Attention Classification

With the rapid development of the Internet, the number of social media and e-commerce platforms increased dramatically. Users from all over world share their comments and sentiments on the Internet become a new tradition. Applying natural language processing technology to analyze the text on the Internet for mining the emotional tendencies has become the main way in the social public opinion monitoring and the after-sale feedback of manufactory. Thus, the study on text sentiment analysis has shown important social significance and commercial value. Sentiment analysis is a hot research topic in the field of natural language processing and data mining in recent ten years. The paper starts with the topic of "Sentiment Analysis using a CNN-BiLSTM deep model based on attention mechanism classification". First, it conducts an in-depth investigation on the current research status and commonly used algorithms at home and abroad, and briefly introduces and analyzes the current mainstream sentiment analysis methods. As a direction of machine learning, deep learning has become a hot research topic in emotion classification in the field of natural language processing. This paper uses deep learning models to study the sentiment classification problem of short and long text sentiment classification tasks. The main research contents are as follows. Firstly, Traditional neural network based short text classification algorithms for sentiment classification is easy to find the errors. The feature dimension is too high, and the feature information of the pool layer is lost, which leads to the loss of the details of the emotion vocabulary. To solve this problem, the Word Vector Model (Word2vec), Bidirectional Long-term and Short-term Memory networks (BiLSTM) and convolutional neural network (CNN) are combined in Quora dataset. The experiment shows that the accuracy of CNN-BiLSTM model associated with Word2vec word embedding achieved 91.48%. This proves that the hybrid network model performs better than the single structure neural network in short text. Convolutional neural network (CNN) models use convolutional layers and maximum pooling or max-overtime pooling layers to extract higher-level features, while LSTM models can capture long- term dependencies between words hence are better used for text classification. However, even with the hybrid approach that leverages the powers of these two deep-learning models, the number of features to remember for classification remains huge, hence hindering the training process. Secondly, we propose an attention based CNN-BiLSTM hybrid model that capitalize on the advantages of LSTM and CNN with an additional attention mechanism in IMDB movie reviews dataset. In the experiment, under the control of single variable of Data volume and Epoch, the proposed hybrid model was compared with the results of various indicators including recall, precision, F1 score and accuracy of CNN, LSTM and CNN-LSTM in long text. When the data size was 13 k, the proposed model had the highest accuracy at 0.908, and the F1 score also showed the highest performance at 0.883. When the epoch value for obtaining the optimal accuracy of each model was 10 for CNN, 14 for LSTM, 5 for MLP and 15 epochs for CNN-LSTM, which took the longest learning time. The F1 score also showed the best performance of the proposed model at 0.906, and accuracy of the proposed model was the highest at 0.929. Finally, the experimental results show that the bidirectional long- and short-term memory convolutional neural network (BiLSTM-CNN) model based on attention mechanism can effectively improve the performance of sentiment classification of data sets when processing long-text sentiment classification tasks. Keywords: sentiment analysis, CNN, BiLSTM, attention mechanism, text classification

Design of Sentiment Analysis Framework of Digital Media Short Text Based on Multi-pattern Sentiment Lexicon

Along the continuous advancement of the network and the rise of digital media, the amount of data produced by the exponential explosion. And how to use these data to provide personalized services for users is one of the current research focuses. To address the issue of insufficient coverage in the current sentiment lexicon and the difficulty of constructing sentiment lexicon in specific fields, this study proposes a multi-modal emotional thesaurus. Semi-supervised learning is used to solve the problem of insufficient coverage of emotional thesaurus, and a semi-supervised classification algorithm is realized by using a large number of unlabeled sample data combined with a small number of labeled sample data. Optimized learning is used to solve the problem of difficult construction of emotional thesaurus in specific fields, the corresponding specific emotional thesaurus is constructed by adaptive adjustment of emotional word score, and finally the improved emotional thesaurus is used to build a digital media short text sentiment analysis framework. For testing, the NLPCC dataset was used in this study, Experiments show that the framework constructed in this study requires 87 iterations, a Recall value of 0.912, a F1 value of 0.753, and an average accuracy of 83.39%, all of which are better than the sentiment analysis framework without the use of multi-pattern sentiment lexicon. In the simulation experiment, the recognition accuracy reached 85.88%, which was 16.85%, 11.57% and 6.72% higher than the test scenarios using a single emotion thesaurus selected in this study. The above results show that the digital media short-text sentiment analysis framework built in this research based on multi-pattern sentiment lexicon can carry out short-text sentiment analysis more accurately and efficiently, so as to accurately analyze users’ needs and provide customized services precisely.

Comparative analysis for augmented decision-making applications using deep learning models

Now a days decision-making plays a significant role in various applications and several research. For applications such as diseases, intelligent routing systems, and online shopping carts such as e-commerce sites, recommended systems are developed based on sentiment analysis (SA) and take accurate decision-making based on the predictions and analyze the accurate decisions based on the result analysis. When it comes to practical uses, deep learning (DL) has by far been the most popular. DL becomes an indispensable domain for several tasks in science and engineering. It is very difficult to take decisions based on traditional tests in various research areas such as disease prediction, textual sentiment analysis, and risk prediction of autonomous vehicles due to the lack of accuracy and long time for results. To address this, various approaches are proposed to adopt. Decision-making is based on multi-criticism, which is more useful to solve critical issues in making accurate decisions than previous approaches. In this paper, an improved and augmented decision-making deep learning algorithm is discussed and shows the comparison among the various DL algorithms. The performance is calculated according to the parameters.

Collaborative fine-grained interaction learning for image–text sentiment analysis

Investigating interactions between image and text can effectively improve image–text sentiment analysis, but most existing methods do not explore image–text interaction at fine-grained level. In this paper, we propose a Memory-enhanced Collaborative Fine-grained Interaction Transformer (MCFIT) to learn collaborative fine-grained interaction between image and text. Specifically, a multi-branch encoder is designed to learn both fine-grained region-word and patch-word interactions. Meanwhile, Memory-enhanced Cross-Attention (MECA) is proposed to utilize patch and region information to improve region-word interaction and patch-word interaction learning, respectively. Therefore, collaborative fine-grained interaction can yield more accurate image–text interaction. Finally, to analyze the sentiments embedded in real-life Chinese image–text pairs, we build a large-scale Chinese image–text sentiment dataset (CISD) containing 54,931 image–text pairs. Extensive experiments conducted on four real-life datasets prove the effectiveness of collaborative fine-grained interaction and demonstrate that MCFIT outperforms the state-of-the-art baselines.

Iraqi Sentiment and Emotion Analysis Using Deep Learning

Analyzing sentiment and emotions in Arabic texts on social networking sites has gained wide interest from researchers. It has been an active research topic in recent years due to its importance in analyzing reviewers' opinions. The Iraqi dialect is one of the Arabic dialects used in social networking sites, characterized by its complexity and, therefore, the difficulty of analyzing sentiment. This work presents a hybrid deep learning model consisting of a Convolution Neural Network (CNN) and the Gated Recurrent Units (GRU) to analyze sentiment and emotions in Iraqi texts. Three Iraqi datasets (Iraqi Arab Emotions Data Set (IAEDS), Annotated Corpus of Mesopotamian-Iraqi Dialect (ACMID), and Iraqi Arabic Dataset (IAD)) collected from Facebook are used to evaluate the model. Experiments showed that the model obtained good results, as the accuracy of the model was 91.1, 92.4, and 92.5% for IADS, ACMID, and IAD, respectively. The results of the model outperformed previous works for all datasets.

Intelligent Task Scheduling Approach for IoT Integrated Healthcare Cyber Physical Systems

Cyber-physical systems (CPS) based on cloud computing provides resources over the Internet and allow a variety of applications to be deployed to provide services for various industries. We proposed IoT-based healthcare cyber-physical system that provides effective resource utilization at fog and cloud levels with minimum execution cost. In addition, we also consider data from social media networking and drug review for the analysis. Furthermore, two different feature extraction approaches were applied based on data collection. Homogeneity score-based K-means clustering is used as a feature extraction and selection method for sensor data features, while text mining and sentiment analysis approach is used for social media networking and drug review data feature extraction. We proposed efficient resource utilization and cost-effective task scheduling at the Fog level and multi-objective heuristic approach Ant colony optimization task scheduling (MOHACO-TS) at cloud level. Both task scheduling algorithms focus on executing maximum task tasks in minimum time with effective resource utilization. We consider five different datasets and existing task scheduling and classification approaches for performance evaluation of the proposed IoT-HCPS framework. From the results, it is evident that the proposed work IoT-HCPS outperformed the exisitng techniques and algorithms.

Sentiment analysis in multilingual context: Comparative analysis of machine learning and hybrid deep learning models

This research paper investigates the efficacy of various machine learning models, including deep learning and hybrid models, for text classification in the English and Bangla languages. The study focuses on sentiment analysis of comments from a popular Bengali e-commerce site, "DARAZ," which comprises both Bangla and translated English reviews. The primary objective of this study is to conduct a comparative analysis of various models, evaluating their efficacy in the domain of sentiment analysis. The research methodology includes implementing seven machine learning models and deep learning models, such as Long Short-Term Memory (LSTM), Bidirectional LSTM (Bi-LSTM), Convolutional 1D (Conv1D), and a combined Conv1D-LSTM. Preprocessing techniques are applied to a modified text set to enhance model accuracy. The major conclusion of the study is that Support Vector Machine (SVM) models exhibit superior performance compared to other models, achieving an accuracy of 82.56% for English text sentiment analysis and 86.43% for Bangla text sentiment analysis using the porter stemming algorithm. Additionally, the Bi-LSTM Based Model demonstrates the best performance among the deep learning models, achieving an accuracy of 78.10% for English text and 83.72% for Bangla text using porter stemming. This study signifies significant progress in natural language processing research, particularly for Bangla, by enhancing improved text classification models and methodologies. The results of this research make a significant contribution to the field of sentiment analysis and offer valuable insights for future research and practical applications.