Earticle

Online reviews, as a notable form of electronic word-of-mouth (eWOM), have demonstrated a significant association with a broad range of consumers, particularly in shaping their impressions of products or services and influencing their purchasing intentions. Both review ratings and sentiment reflect the reviewer’s satisfaction and attitude after using a product or service. Additionally, reviewers with greater relevant experience are often able to provide more insightful evaluations. Based on this, our study analyzes the relationship between review ratings and textual sentiment and their link to review helpfulness, using movie reviews collected from IMDb.com, while also considering the moderating role of the reviewer’s relevant experience. Our findings indicate that extreme ratings are more strongly related to review helpfulness. Moreover, regardless of the sentiment expressed in the review title and text, reviews with more negative sentiment are generally perceived as more helpful. Reviewer experience, especially when relevant to the movie genre being reviewed, amplifies the connection between rating, title sentiment, and text sentiment and review helpfulness. This study confirms the relationship between three key factors and review helpfulness, further clarifying the moderating role of reviewer relevant experience. These results contribute to eWOM research and enrich the literature on review helpfulness.

The rapid growth of generative AI has raised concerns about content authenticity on user-generated platforms, particularly in online reviews. This study proposes an interpretable, feature-based machine learning approach to detect AI-generated reviews, focusing on transparency and efficiency. By integrating linguistic feature analysis (LIWC), textual pattern recognition (TF-IDF), and Large Language Model (LLM)-based interpretation, Random Forest and XGBoost classifiers were applied to achieve robust predictive performance. SHAP value analysis was used to enhance interpretability by identifying key linguistic and structural patterns distinguishing AI-generated content from human-written reviews. The findings reveal that AI-generated reviews tend to exhibit structured grammar, formulaic conclusions, exaggerated sentiment, and broader aspect coverage compared to the nuanced and informal style of human reviews. This study contributes to the field by offering (1) an effective feature-based detection framework, (2) empirical validation of linguistic distinctions between AI and human content, and (3) practical guidance for developing lightweight, trustworthy AI-content detection tools.

This study demonstrates how online hotel reviews can classify customer loyalty by analyzing textual features such as sentiment, rating, and loyalty keywords. Using a review dataset from TripAdvisor, the research applies sentiment analysis tools (VADER, TextBlob, SenticNet) and topic modeling techniques to classify loyalty. Results show that loyalty keywords imply a significant difference in their presence between loyal and non-loyal customers. At the same time, a considerable difference in their presence between loyal and non-loyal customers, while sentiment scores present significantly moderate results. However, Rao-Striling diversity and review length do not exhibit a significant difference. The study contributes a framework for using review content to identify loyal customers, offering practical methods for businesses to enhance customer retention and improve CRM strategies. Limitations include reliance on sentiment scores for loyalty classification and the exclusion of behavioral data, suggesting future research could incorporate more comprehensive customer data for validation.

With the advancements and widespread adoption of social media platforms, electronic word-of-mouth (eWOM) has become increasingly diverse, significantly impacting consumer decisions through its varied information dissemination. Compared to textual reviews, video reviews on the social medias are becoming more crucial in consumers' shopping decision-making process. This is attributed to video reviews offering consumers more comprehensive information through visual cues and dynamic product explanations. While the perceived helpfulness of information presented in video reviews can reflect viewers' evaluations of review quality, there is a notable absence of research examining the factors influencing the assessment of review helpfulness in videos. To bridge this gap, we investigated the impact of YouTube channel characteristics and video duration on the helpfulness of video reviews for electronic products. Additionally, this study applied speech-to-text conversion and text mining techniques to extract and analyze emotional factors. The findings of the model analysis revealed that a greater number of subscribers to a channel corresponds to increased helpfulness of video reviews. Furthermore, the helpfulness of video reviews tends to rise when the expressed emotions are more negative. This study also confirmed a non-linear relationship between video duration and review helpfulness, concluding that videos of an appropriate duration are most beneficial to consumers. This study not only adapted and validated the antecedents of helpfulness from textual reviews to video reviews but also expanded research on the review helpfulness within the realm of social media. Drawing on the analytical outcomes, practical insights were provided for the generation and management of eWOM and video reviews on social media platforms. Especially, the results provide constructive for individuals and platforms involved in generating and presenting more helpful content for product reviews.

유용한 온라인 리뷰는 고객에게 가치 있는 정보를 제공하고 불확실성을 줄여주며 구매 의사결정을 지원한다. 온라인 리뷰가 제공하는 풍부한 정보는 리뷰 유용성에 영향을 미치는 요인으로 리뷰 유용성 간의 영향관계를 살펴볼 필요가 있다. 따라서 본 연구에서는 온라인 리뷰가 제공하는 정보적 요인이 리뷰 유용성에 미치는 영향관계를 살펴보고 제품의 유형에 따른 조절효과를 확인하고자 한다. 본 연구에서는 JD.com의 온라인 리뷰를 제품유형에 따라 수집하였으며, 텍스트 길이와 이미지 수가 리뷰 유용성에 미치는 영향이 역 U자 형의 비선형 관계를 나타냄을 확인하였다. 또한 동영상이 리뷰 유용성에 영향을 미치며, 정보의 풍부함이 리뷰 유용성에 긍정적 영향을 미친다는 사실을 확인하였다. 제품유형에 따른 조절효과에서는 리뷰의 길이, 이미지, 동영상, 그리고 이 세 가지 정보 구성 요인이 제공하는 정보 풍부성과 리뷰 유용성 간의 관계에서 탐색제품보다 경험제품이 더 강하게 나타났다. 분석을 통해 본 연구는 온라인 리뷰의 다양한 콘텐츠 요소가 정보처리에 미치는 영향과 온라인 쇼핑에서의 의사결정에 도움이 되는 방식을 확인하였다. 이러한 연구결과는 리뷰 유용성 평가에 있어 온라인 리뷰 구조의 역할에 대한 이해를 심화시키며, 플랫폼이 리뷰 생성 및 추천을 위한 효과적인 전략을 수립하고, 리뷰어가 잠재고객에게 더 유용한 리뷰를 작성하는 데 기여할 수 있도록 정보를 제공한다.

Helpful online reviews provide customers with valuable information, reduce uncertainties, and support purchasing decisions. Although previous studies have identified various factors that influence review helpfulness, the relationship between informational elements and review helpfulness requires further analysis and clarification. This study analyzed online shopping review, revealing that the impact of text length and number of images on review helpfulness is nonlinear, exhibiting an inverted U-shaped effect. Additionally, our study confirmed that videos significantly enhanced perceived helpfulness and that the information richness of a review positively influenced its helpfulness. The moderating effect of product type shows that the relationship between review length, images, videos, and the richness of information provided by these factors and review helpfulness is stronger for experience goods than for search goods. The analysis explains the influence and significance of various content components in shaping online reviews and their influence on aiding information processing for online shopping. These findings deepen our understanding of the role of online review structures in evaluating review helpfulness, assist platforms in formulating more effective strategies for review generation and recommendations, and guide reviewers in writing reviews that are more beneficial to potential customers.

ChatGPT has had a significant impact on society, business, and academia by influencing individuals and organizations through knowledge generation and supporting users in locating conversational inquiries and answers. It can transform how people seek answers by combining human-like conversational skills with AI. By eradicating the cumbersome process of selecting from multiple options, users can conduct preliminary research or create optimized solutions. The purpose of this research is to investigate how consumers use ChatGPT and digital transformation, specifically in terms of knowledge development, searching and recommending, and optimizing accessible possibilities. Using many linked theories, we address the potential implications and insights that can be gained from ChatGPT’s early stages and its integration with other applications such as robotics, service automation, and the metaverse. Finally, the application of ChatGPT has practical, theoretical, and phenomenological impacts, in addition to improving users’ experiences.

In the dynamic world of e-commerce, effectively deciphering customer reviews is of paramount importance. This study uniquely combines the RFE-SHAP feature selection technique with topic modeling (LDA) to address prevalent challenges like overfitting in predictive modeling. Our empirical analysis underscores the superior performance of the Random Forest model, particularly when refined with a subset of 14 pivotal features. Notably, topics such as quality and appearance, fit and comfort, and durability concerns emerged as significant determinants of customer satisfaction within the clothing sector. Utilizing data exclusively from Amazon's clothing reviews, our research emphasizes the criticality of strategic feature selection and delves deep into the multifaceted factors shaping customer sentiments. By seamlessly merging quantitative metrics with qualitative content insights, this study not only offers a robust framework for understanding online reviews but also paves the way for future research in optimizing e-commerce strategies based on customer feedback.

With the widespread use of social media, online social platforms like Twitter have become a place of rapid dissemination of information―both accurate and inaccurate. After the COVID-19 outbreak, the overabundance of fake information and rumours on online social platforms about the COVID-19 pandemic has spread over society as quickly as the virus itself. As a result, fake news poses a significant threat to effective virus response by negatively affecting people’s willingness to follow the proper public health guidelines and protocols, which makes it important to identify fake information from online platforms for the public interest. In this research, we introduce an approach to detect fake news using deep learning techniques, which outperform traditional machine learning techniques with a 93.1% accuracy. We then investigate the content differences between real and fake news by applying topic modeling and linguistic analysis. Our results show that topics on Politics and Government services are most common in fake news. In addition, we found that fake news has lower analytic and authenticity scores than real news. With the findings, we discuss important academic and practical implications of the study.

The detection of products with manipulated reviews has received widespread research attention, given that a truthful, informative, and useful review helps to significantly lower the search effort and cost for potential customers. This study proposes a method to recognize products with manipulated online customer reviews by examining the sequence of each review’s sentiment, readability, and rating scores by product on randomness, considering the example of a Russian online retail site. Additionally, this study aims to examine the association between brand awareness and existing manipulation with products’ reviews. Therefore, we investigated the difference between well-known and unknown brands’ products online reviews with and without manipulated reviews based on the average star rating and the extremely positive sentiment scores. Consequently, machine learning techniques for predicting products are tested with manipulated reviews to determine a more useful one. It was found that about 20% of all product reviews are manipulated. Among the products with manipulated reviews, 44% are products of well-known brands, and 56% from unknown brands, with the highest prediction performance on deep neural network.

In the post-COVID era, several companies are struggling. The rapid rise of COVID-19 from the beginning of 2020 has led to the situation in which most of our society around the world has to work from home, maintain social distance, and live. Therefore, SMEs affected by COVID-19 are trying to break away from the existing CRM to communicate with customers smoothly and use social media, the most powerful marketing tool of the present era, to explore the process of adopting social CRM. The research purpose of this paper is to find out how much COVID-19 has influenced Social CRM (sCRM) adoption in SMEs and what variables have influenced sCRM adoption, more specifically with TOE research theory variables.

In the context of the COVID-19 pandemic Information and communication technologies (ICTs) have emerged as the critical enabler, almost all social activity and shopping is done online. Therefore, online customer reviews (OCRs) have a great impact on customers' purchase decision-making process. Since some companies or sellers strategically create fake online reviews in an effort to influence customers' purchase decisions, the detection of fake (deception) plays a critical role in e-commerce. However, researchers in fake review detection faced a lot of problems. One of them is the lack of a high-quality fake review dataset. The purpose of our research is to study fake reviews' features based on a publicly available dataset and test different machine learning methods for detecting fake reviews through analyzing sentiments and readability.

The WeChat mini program is an application embedded in WeChat that users can use without downloading and installing. After it was officially released in 2017, many travel enterprises have launched their own mini programs on the WeChat platform. This study applies affordance theory to investigate the WeChat mini program’s role in tourism activities through social network analysis using the R programming language. The authors searched the topic of “how do you perceive the travel-related WeChat mini program” and then crawled the 200 comments found; 180 comments were analyzed after data cleansing. The results show that travel-related WeChat mini programs play a very important role in Chinese social network tourism activities. This paper found that WeChat played a more active role in various tourism-related interactions with Chinese social networks. Moreover, the results show how affordance theory is applied to the use of WeChat mini programs.

The purpose of this study is to explore the effect of different types of simultaneous presentation (i.e., reviewer information, textual and visual content, and similarity between textual-visual contents) on review usefulness and review enjoyment in online restaurant reviews (ORRs), as they are interrelated yet have rarely been examined together in previous research. By using Latent Dirichlet Allocation (LDA) topic modeling and state-of-the-art machine learning (ML) methodologies, we found that review readability in textual content and salient objects in images in visual content have a significant impact on both review usefulness and review enjoyment. Moreover, similarity between textual-visual contents was found to be a major factor in determining review usefulness but not review enjoyment. As for reviewer information, reputation, expertise, and location of residence, these were found to be significantly related to review enjoyment. This study contributes to the body of knowledge on ORRs and provides valuable implications for general users and managers in the hospitality and tourism industries.

Although it has been proven that textual features of review content have an impact on review usefulness/enjoyment, visual features of review content have been paid less attention. We therefore identify three major photographic components of images (i.e., color, composition, and image category) embedded in an online restaurant review (ORR)and investigate their influence on both review usefulness and review enjoyment. Color is measured by calculating pixel intensities of a photo, composition is measured by analyzing a photo using the role of thirds, and image category is classified by employing the Google Vision API. This study is expected to build a new layer on the literature on ORRs by suggesting a new way to identify and measure photographic components, combining two research areas of photography and image mining. Practically, research findings will suggest a way to make more useful/enjoyable ORRs for general users and provide practical implications for managers in the hospitality and tourism industry.

Drawing on the elaboration likelihood model, this study explores the antecedents of perceived review helpfulness focused on the argument quality consisting of emotional persuasion, topic relevance, and comprehensiveness. Controlling for the effect of source credibility, this study has found that three negative emotions significantly associated with perceived review helpfulness. Fear and sadness are found to positively affect perceived helpfulness with more emotional appeals. Also, this study highlighted the importance of topic relevance in information persuasion. Issue-relevant arguments could motivate one to scrutinize the details of a review and thus perceive it as helpful. Our study contributes to provide a theoretical basis for future study by including emotion as a predominant element of argument quality in the elaboration likelihood model. The findings could be applicable to further study on negativity bias within the online context. The findings also carry managerial implications that are discussed.

Online reviews provide an increasingly important role in the purchase decision making. Many customers prefer to explore helpful reviews in order to diminish their searching effort such as time effort. This paper examines the differential effects of three discrete emotions (anger, fear, sadness) embedded in a customer review on the perceived review helpfulness. We analyze “verified purchase” reviews from Amazon.com to examine the relationship between emotions and review helpfulness. The findings of this study extend previous research by suggesting that product type moderates the effects of emotions on the helpfulness of the review. Anger embedded in a customer review has a greater negative on the perceived review helpfulness for experience goods than for search goods. Fear and sadness are found to have a positive and negative effect on the perceived review helpfulness respectively, which shows no difference between search goods and experience goods. The implications of these findings contribute to a better understanding of the important role of emotions embedded in reviews on the perceived review helpfulness. The findings also provide practical insights (i.e., the presentation of online reviews) for the retailers or review platforms, and give suggestions for the consumers on how to select and write a helpful review.

Many studies have focused on the prediction of corporate credit rating using various data mining techniques. One of the most frequently used algorithms is support vector machines (SVM), and recently, novel techniques such as SVM+ and SVM+MTL have emerged. This paper intends to show the applicability of such new techniques to multi-classification and corporate credit rating and compare them with conventional SVM regarding prediction performance. We solve multi-class SVM+ and SVM+MTL problems by constructing several binary classifiers. Furthermore, to demonstrate the robustness and outstanding performance of SVM+MTL algorithm over other techniques, we utilized four typical multi-class processing methods in our experiments. The results show that SVM+MTL outperforms both conventional SVM and novel SVM+ in predicting corporate credit rating. This study contributes to the literature by showing the applicability of new techniques such as SVM+ and SVM+MTL and the outperformance of SVM+MTL over conventional techniques. Thus, this study enriches solving techniques for addressing multi-class problems such as corporate credit rating prediction.

SNS의 확산으로 온라인 상점에서는 상품에 대한 주관적인 의견이 내포되어 있는 고객리뷰 정보가 빠르게 생성되고 확산되어 다른 고객들에게 큰 영향을 미치고 있다. 이와 더불어, 고객들의 긍정적 또는 부정적 의견을 분석하여 개선방안을 모색하려는 오피니언마이닝(opinion mining)이 주목 받고 있다. 고객리뷰에 내포된 감성정보를 가진 용어들은 감성분류를 하는데 가장 중요한 역할을 하기 때문에 영향력이 높은 용어를 선별하는 것이 가장 중요하다. 본 연구에서는 품사태깅을 이용하여 최적의 용어들을 선별하고 용어정보에 기반한 문서수준에서의 감성분류모형을 제안하고자 한다. 고객 리뷰의 감성분류모형에 대표적인 기계학습기법인 SVM을 적용하고, SVM의 입력변수 선정과정에 품 사태깅 방식과 용어추출기법을 다르게 조합하고 사용하여 긍정적/부정적 문서를 분류하였다. 본 연구 에서 제안한 감성분류모형의 성과를 검증하기 위해 아마존(Amazon.com)의 영화와 도서에 대한 고객리 뷰 80,000개를 수집하여 불필요한 용어들을 제거한 후 품사태깅을 통해 용어를 추출하였다. 추출된 용어는 문서빈도, TF-IDF, 정보획득량, 카이제곱 통계량의 값을 산출하여 값을 통해 용어들을 순위화하 고, 각 상위 20개에 해당하는 최적의 용어를 선정한 후 SVM을 이용하였다. 제안된 감성분류모형을 통해 기존 연구에서 언급한 형용사만을 사용한 예측변수와 4품사를 사용한 예측변수에서의 실험결과 를 통해 비교 분석하였다. 카이제곱 통계량 기반의 감성분류모형이 다른 모형보다 예측성과가 가장 우수하게 나타나는 것을 확인할 수 있었다. 본 연구에서 제안된 문서수준에서의 용어기반 감성분류모 형을 이용함으로써 온라인 상점에서의 서비스 개선과 경쟁력 확보에 많은 도움이 될 것으로 기대된다.

Customer reviews which include subjective opinions for the product or service in online store have been generated rapidly and their influence on customers has become immense due to the widespread usage of SNS. In addition, a number of studies have focused on opinion mining to analyze the positive and negative opinions and get a better solution for customer support and sales. It is very important to select the key terms which reflected the customers’ sentiment on the reviews for opinion mining. We proposed a document-level terms-based sentiment classification model by select in the optimal terms with part of speech tag. SVMs (Support vector machines) are utilized to build a predictor for opinion mining and we used the combination of POS tag and four terms extraction methods for the feature selection of SVM. To validate the proposed opinion mining model, we applied it to the customer reviews on Amazon. We eliminated the unmeaning terms known as the stopwords and extracted the useful terms by using part of speech tagging approach after crawling 80,000 reviews. The extracted terms gained from document frequency, TF-IDF, information gain, chi-squared statistic were ranked and 20 ranked terms were used to the feature of SVM model. Our experimental results show that the performance of SVM model with four POS tags is superior to the benchmarked model, which are built by extracting only adjective terms. In addition, the SVM model based on Chi-squared statistic for opinion mining shows the most superior performance among SVM models with 4 different kinds of terms extraction method. Our proposed opinion mining model is expected to improve customer service and gain competitive advantage in online store.

Support vector machines (SVMs), a machine learning technique, has been applied to not only binary classification problems such as bankruptcy prediction but also multi-class problems such as corporate credit ratings. However, in general, the performance of SVMs can be easily worse than the best alternative model to SVMs according to the selection of predictors, even though SVMs has the distinguishing feature of successfully classifying and predicting in a lot of dichotomous or multi-class problems. For overcoming the weakness of SVMs, this study has proposed an approach for selecting features for multi-class SVMs that utilize the impurity measures of classification trees. For the selection of the input features, we employed the C4.5 and CART algorithms, including the stepwise method of discriminant analysis, which is a well-known method for selecting features. We have built a multi-class SVMs model for credit rating using the above method and presented experimental results with data regarding S&P 500 companies.

정보기술의 발전과 더불어 기업과 고객간의 대부분의 정보가 축적되면서 기업은 거래고객의 자세한 정보를 활용하여 차별화된 마케팅을 제공할 수 있다. 본 연구는 기업이 제공하는 마케팅 전략을 보다 효과적으로 실행하기 위해 고객을 세분화하고, 세분화된 고객집단별 마케팅 프로모션에 대한 반응을 예측하는 모형을 제시하였다. 고객세분화에는 데이터마이닝 기법 중 SOM(Self-organizing Map)을 적용하였으며, 세분화된 집단별 프로모션 반응예측에는 로짓모형, 신경망 등의 단일모형과 k-최근접이웃법을 이용한 단일모형들의 통합모형을 적용하였다. 제시된 방법론으로 기업은 프로모션에 대한 고객반응을 예측할 뿐만 아니라 프로모션에 대한 반응을 쉽게 예측할 수 있는 고객집단과 반응예측이 어려운 고객집단으로 구분하여 프로모션의 효과를 극대화하고 각 집단에 맞는 프로모션 전략을 수립할 수 있다.

This paper proposed a method that segmented customers utilizing SOM(Self-organizing Map) and predicted the customers' response of a marketing promotion for each customer's segments. Our proposed method focused on predicting the response of customers dividing into customers' segment whereas most studies have predicted the response of customers all at once. We deployed logistic regression, neural networks, and support vector machines to predict customers' response that is a kind of dichotomous classification while the integrated approach was utilized to improve the performance of the prediction model. Sample data including 45 variables regarding demographic data about 600 customers, transaction data, and promotion activities were applied to the proposed method presenting classification matrix and the comparative analyses of each data mining techniques. We could draw some significant promotion strategies for segmented customers applying our proposed method to sample data.