Preference mining and fuzzy inference for hotel selection based on aspect-based sentiment analysis from user-generated content

A Pre-Trained Model for Aspect-based Sentiment Analysis Task: using Online Social Networking

Due to the rapid increase in User-Generated Content (UGC) data, opinion mining, also called sentiment analysis, has attracted much attention in both academia and industry. Aspect-Based Sentiment Analysis (ABSA), a subfield of sentiment analysis, aims to extract the aspect and the corresponding sentiment simultaneously. Previous works in ABSA may generate undesired aspects, require a large amount of training data, or produce unsatisfactory results. This paper proposes a Graph Neural Network based method to automatically generate aspect-specific sentiment words using a small number of aspect seed words and general sentiment words. It subsequently leverages the aspect-specific sentiment words to improve the Joint Aspect-Sentiment Autoencoder (JASA) model. We conduct experiments on two datasets to verify the proposed model. It shows that our approach has better performance in the ABSA task when compared with previous works.KeywordsAspect-based sentiment analysisAttention mechanismAutoencoderText graph convolution networkText mining

Social networking sites are now entered into the mainstream media landscape and cover a broad demographic spectrum of Internet users. Users post real time messages about their opinions on different topics and current issues, complain and express sentiments for products they use in their daily life. Such users have produced extensive online user-generated content and attracted business firms to consider it as a valuable input for their business processes. But this rapid growth in data size has posted many challenges in accessing, managing, analyzing, and integrating the data in networking, management, computational, and statistical sciences. This paper proposes a framework for analyzing the online data produced by the social networking site, twitter. The proposed Jen-Ton framework collects, pre-processes and analyses the sentiment present in the tweets and prepares the aspect based sentiment analysis for a product. This framework uses three techniques namely, Imputation of Missing Sentiment (IMS), Aspect based Sentiment Analysis using Fuzzy Logic (ASFuL) and Aspect based SenTiment Analysis using Clustering with Genetic Algorithm (ASTA-CGA) for enhancing the accuracy of the aspect based sentiment analysis in big data.

Aspect-based sentiment analysis (ABSA) aims to identify the sentiment polarity associated with specific aspect terms within sentences. Existing studies have primarily focused on constructing graphs from dependency trees of sentences to extract syntactic features. However, given that public datasets are often derived from online reviews, the syntactic structures of these sentences frequently exhibit irregularities. As a result, the performance of syntactic-based Graph Convolution Network (GCN) models is adversely impacted by the noise introduced during dependency parsing. Moreover, the interaction between syntactic and semantic information in these approaches is often insufficient, which significantly impairs the model's ability to accurately detect sentiment.To address these challenges, we propose a novel approach called Syntactic Denoising with Multi-strategy Auxiliary Enhancement (SDMAE) for the ABSA task. Specifically, we prune the original dependency tree by focusing on context words with specific part-of-speech features that are critical for conveying the sentiment of aspect terms, and then construct the graph. We introduce a Multi-channel Adaptive Aggregation Module (MAAM), a feature aggregation system that employs a multi-head attention mechanism to integrate semantic and syntactic GCN output representations. Furthermore, we design a multi-strategy task learning framework that incorporates sentiment lexicons and supervised contrastive learning to enhance the model's performance in aspect sentiment recognition.Comprehensive experiments conducted on four benchmark datasets demonstrate that our approach achieves significant performance improvements compared to several state-of-the-art methods across all evaluated datasets.

Aspect-based opinion mining is one among the thought-provoking research field which focuses on the extraction of vivacious aspects from opinionated texts and polarity value associated with these. The principal aim here is to identify user sentiments about specific features of a product or service rather than overall polarity. This fine-grained polarity identification about myriad aspects of an entity is highly beneficial for individuals or business organizations. Extricating these implicit or explicit aspects can be very challenging and this paper elaborates copious aspect extraction techniques, which is decisive for aspect-based sentiment analysis. This paper presents a novel idea of combining several approaches like Part of Speech tagging, dependency parsing, word embedding, and deep learning to enrich the aspect-based sentiment analysis specially designed for Twitter data. The results show that combining deep learning with traditional techniques can produce excellent results than lexicon-based methods.

Aspect-based sentiment analysis currently attracts much attention from researchers in sentiment analysis and opinion mining fields. In this problem we simultaneously solve both tasks of aspect detection and sentiment detection. This paper proposes a Convolutional Neural Network based model in which we integrate extended rich information features into the basic CNN model. Our experiment is conducted on the aspect-based sentiment analysis task of Semeval 2016 and achieves the best results in comparison with previous studies.

The fuzzy logic inference system was developed to identify driving intention. The membership functions and rules of the fuzzy logic inference system were built by using mathematical statistics and neural network. The vehicle model was built based on a series–parallel hybrid vehicle using Cruise software. The driving intention inference system was designed in Simulink. The simulation is done based on Cruise and Simulink. The simulation results prove that the fuzzy inference system can identify driving intentions excellently and the control strategy based on driving intentions can help to reduce more fuel consumption.KeywordsDriving intentionFuzzy inferenceHybrid vehiclesNeural networkSimulation

In the ever-expanding digital landscape, the abundance of user-generated content on consumer platforms such as Booking and TripAdvisor offers a rich source of information for both travellers and hoteliers. Sentiment analysis, a fundamental research task of natural language processing (NLP) is used for mining sentiments and opinions within this vast reservoir of text reviews. A more specific type of sentiment analysis, i.e., aspect-based sentiment analysis (ABSA), is used when processing customer reviews is required. In ABSA, we aim to capture aspect-level sentiments and intricate relationships between various aspects within reviews. This article proposes a novel approach to ABSA by introducing a novel technique of word sense disambiguation (WSD) and integrating it with the Transformer architecture bidirectional encoder representations from Transformers (BERT) and graph convolutional networks (GCNs). The proposed approach resolves the intriguing ambiguities of the words and represents the review data as a complex graph structure, facilitating the modeling of intricate relationships between different aspects. The combination of bidirectional long short-term memory (BiLSTM) and GCN proves effective in capturing inter-dependencies among various aspects, providing a nuanced understanding of customer sentiments. The experiments are conducted on the RABSA dataset (an enhanced and richer hotel review data collection), and results demonstrate that our approach outperforms previous baselines, showcasing the effectiveness of integrating WSD in ABSA. Furthermore, an ablation study confirms the significant contribution of the WSD module to the overall performance. Moreover, we explore different similarity measures and find that cosine similarity yields the best results when identifying the real sense of a word in a given sentence using WordNet. The findings of our work and future work related to our work create lots of interest for people in the tourism and hospitality industry. This research gives another boost to the concept of the potential of NLP techniques in sentiment analysis. It emphasizes that if we combine the potential of NLP techniques along with state-of-the-art machine learning frameworks, we can shape the future of this field.

Semantics perception and refinement network for aspect-based sentiment analysis

Aspect-Based Sentiment Analysis (ABSA) plays a vital role in interpreting user opinions in customer reviews by identifying sentiment polarity linked to specific service aspects. However, traditional ABSA methods largely focus on explicit aspect mentions, often neglecting implicit aspects that are contextually implied. This limitation is especially critical in the hospitality domain, where customer reviews frequently include indirect references to hotel features. This study proposes an implicit aspect analysis framework for sentiment classification using hotel reviews, aiming to improve the detection and interpretation of implicit aspect expressions. A dataset of 2,000 TripAdvisor hotel reviews was collected and manually annotated by linguistic experts for four aspects: staff service, cleanliness, value for money, and location convenience. Each sentence was labeled as either explicitly or implicitly expressing an aspect, with associated sentiment polarity (positive or negative). A two-stage framework was developed: the first stage employed Word2Vec Skip-gram modeling and TF-IDF to construct an implicit aspect corpus; the second stage implemented machine learning and deep learning classifiers—including SVM, LR, MNB, RF, CNN, and LSTM—to predict sentiment. The hybrid feature representation combining TF-IDF and Word2Vec significantly enhanced the model's ability to detect implicit sentiments. Deep learning models, particularly LSTM and CNN, outperformed traditional classifiers in both sentiment polarity classification and implicit aspect association. Clustering results confirmed strong alignment with expert-labeled data, validating the framework’s effectiveness. The proposed framework demonstrates substantial improvements in classifying implicit aspects, offering valuable insights for service quality evaluation in hospitality and other domains relying on user-generated content.

A number of businesses, such as TripAdvisor, Open Table, and Yelp, have successfully utilized aspect-based sentiment analysis in order to gain insights from reviews provided by customers and enhance the quality of their goods or services. Businesses are able to swiftly discover any unfavorable sentiment or possible harm to their brand when they analyze client input across numerous aspects from social media, online reviews, and conversations with customer care representatives. This study aims to explain how aspect-based semantic analysis of market-collected user-generated data through performance comparisons of Doc2vec and TF-IDF vectorization. Both Doc2Vec and TF-IDF have their own distinctive qualities, which might vary according on the nature of the job, the dataset, and the volume of the available training data. For the objectives of this research, the data was obtained from several of fashion merchants that run their companies by means of the Shopee platform, which is a well-known online marketplace platform in Indonesia. In this research, the accuracy and F1 Score achieved by Doc2Vec vectorization was superior to those achieved by TF-IDF vectorization. Our findings shows that Doc2Vec vectorization is better for classifying customer ratings because it can pull out the semantic meaning of words in a document. The findings also shows that the score of c and gamma parameter have significant impact to the score of Accuracy and F1 Score of the classifier.By precisely categorizing client sentiment, this study enables businesses to improve their services, respond to customers' problems, and increase their customer satisfaction.

This study analyzes 24,183 TripAdvisor reviews on six major attractions in Hong Kong, Macao, and Zhuhai to examine cross-regional tourism governance in the Guangdong–Hong Kong–Macao Greater Bay Area. By applying Latent Dirichlet Allocation (LDA) and Aspect-Based Sentiment Analysis (ABSA), the research uncovers concentrated negative perceptions regarding crowd management and service quality, as well as differences in governance characteristics across regions. The findings highlight the value of User-Generated Content (UGC) as evidence for governance and propose measures such as establishing collaborative platforms, early warning mechanism and digital governance tools to enhance governance effectiveness and promote high-quality tourism development.

In the condition of large alignment angles which brought about nonlinear problem in SINS, a precise SINS error model was established on the concept of Euler platform error angles. To reduce the computation of initial alignment in SINS, a Simplified UKF (SUKF) could be used since its state equation was nonlinear while the measurement equation was linear. A novel method combined Fuzzy Inference System (FIS) and system noise's online estimation together on the basis of SUKF was proposed to adjust the system noise covariance, and online improve the performance of the SINS initial alignment. In the SINS alignment for large misalignment angles simulation conditions, the SUKF via FIS showed higher accuracy, better stability and also better real-time performance compared with conventional UKF.

A fuzzy inference system (FIS) was developed to detect obstructive sleep apnea (OSA) by analyzing the respiratory airflow signal in adults. The parameters analyzed were the normalized area and the standard deviation of consecutive 3-second intervals of baseline adjusted and rectified airflow signal. Fuzzy logic was used to process these parameters to detect apnea and hypopnea when the output values were within a specified range extracted from OSA patient data. The FIS comprised of three major stages of computation: fuzzification, fuzzy rule evaluation and defuzzification. Seven males and two females with an average age of 48 years (range: 26 - 66 years), an average weight of 102 kg (range: 63 - 159 kg), an average height of 1.7 m (range: 1.5 - 1.8 m) and an average body mass index (BMI) of 33 kg/m/sup 2/ (range: 21 - 42 kg/m/sup 2/) participated in this study. Patients spent at least 8 hours in an accredited sleep laboratory. However, patient data was collected for only part of this time. The total amount of test time for all nine patients was 38.83 hours with an average of 4.31 hours/patient (range: 1.92 - 7.63 hours). The total number of apnea events occurring during this time was 808, and the number of hypopnea events was 694. The membership functions for the FIS were derived by analyzing apnea and hypopnea events in four patients. The data from all nine patients were used in algorithm performance evaluation. The apnea and hypopnea events were scored by a sleep specialist and were used to test the correct detection rate by the FIS. The results demonstrated that the FIS reached an overall correct detection rate of 83% across all patients. The false negative rate was 17% and the false positive rate was 12%. The correct detection rate varied from patient to patient and correct rates greater than 90% were achieved in three patients. This study suggests that fuzzy inference could provide an intelligent algorithm for control of a continuous positive airway pressure (CPAP) machine. It would detect apnea and hypopnea events and automatically adjust the pressure to eliminate them. The performance of the algorithm could be further optimized to give increased detection rates. This could be achieved by performing further studies on a larger OSA patient population and utilizing augmentative methods such as neural networks to better sense the fuzzy patterns in the OSA data.

Interpolativity of at-least and at-most models of monotone fuzzy rule bases with multiple antecedent variables