• Journal of Information Processing Systems
• /
• v.18 no.5
• /
• pp.650-664
• /
• 2022

Mobile applications can be easily downloaded and installed via markets. However, malware and malicious applications containing unwanted advertisements exist in these application markets. Therefore, smartphone users install applications with reference to the application review to avoid such malicious applications. An application review typically comprises contents for evaluation; however, a false review with a specific purpose can be included. Such false reviews are known as fake reviews, and they can be generated using artificial intelligence (AI)-based text-generating models. Recently, AI-based text-generating models have been developed rapidly and demonstrate high-quality generated texts. Herein, we analyze the features of fake reviews generated from Generative Pre-Training-2 (GPT-2), an AI-based text-generating model and create a model to detect those fake reviews. First, we collect a real human-written application review from Kaggle. Subsequently, we identify features of the fake review using natural language processing and statistical analysis. Next, we generate fake review detection models using five types of machine-learning models trained using identified features. In terms of the performances of the fake review detection models, we achieved average F1-scores of 0.738, 0.723, and 0.730 for the fake review, real review, and overall classifications, respectively.

• The Journal of Information Systems
• /
• v.31 no.4
• /
• pp.01-23
• /
• 2022

Purpose This study aims to propose a deep learning-based fake review detection model by combining authors' linguistic features and semantic information of reviews. Design/methodology/approach This study used 358,071 review data of Yelp to develop fake review detection model. We employed linguistic inquiry and word count (LIWC) to extract 24 linguistic features of authors. Then we used deep learning architectures such as multilayer perceptron(MLP), long short-term memory(LSTM) and transformer to learn linguistic features and semantic features for fake review detection. Findings The results of our study show that detection models using both linguistic and semantic features outperformed other models using single type of features. In addition, this study confirmed that differences in linguistic features between fake reviewer and authentic reviewer are significant. That is, we found that linguistic features complement semantic information of reviews and further enhance predictive power of fake detection model.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2014.05a
• /
• pp.79-81
• /
• 2014

People increasingly rely on sources of information from E-commerce reviews. Product reviews is an important determinant of potential customers' buying choices. They are also utilized by product manufacturers to find problems of their products and to collect competitive intelligence information about their competitors. Unfortunately, it is well-known that many online product reviews are not made by genuine costumers of products. Reviewers could write some undeserving positive reviews to promote or fake negative reviews to defame some certain product, and we call them fake product reviews. Fake product review detection makes an attempt to detect fake reviews and removes them to restore the truthful ones for readers. To the best of our knowledge, there is still less published study on this problem. In this paper, we make a survey and an attempt to give a brief overview on fake product review detection. The related work of fake product review detection is presented including web spam and spam email. Then some methods to detect fake reviews are introduced and summarized. The trend of fake product review detection is concluded finally.

• Journal of Fashion Business
• /
• v.21 no.3
• /
• pp.43-57
• /
• 2017

This study concerns visual humor in fake technique fashion. While previous studies focused mainly on expression techniques of fake technique fashion, this study analyzed visual humor in fake technique fashion based on classification criteria of visual humor expression techniques, differenting this study from other studies. The purpose of this study was to derive visual humor in fake technique fashion by classifying cases of fake technique fashion, and re-classifying outcomes of primary classification based on criteria of visual humor expression techniques. As for methods, this theoretical study was conducted on humor, expression techniques of visual humor, fake fashion and fake expression techniques through literature review. Subsequently, 485 fake technique fashion images obtained from research were classified by expression techniques, and cases of fake technique fashion were analyzed. In addition, by combining this theoretical study with case studies, fake technique fashion was re-classified according to criteria of visual humor expression techniques to derive the characteristics of visual humor in fake technique fashion. Based on visual humor expression techniques, visual humor in fake technique fashion was created by distortion and transformation that made the fake look real by distorting or transforming the fake, enlargement and reduction that created new forms by altering familiar forms, and typeplay that added fun by changing familiar luxury logos into various forms.

• Information Systems Review
• /
• v.23 no.1
• /
• pp.45-68
• /
• 2021

As the influence of online reviews on consumer decision-making increases, concerns about review manipulation are also increasing. Fake reviews or review manipulations are emerging as an important problem by posting untrue reviews in order to increase sales volume, causing the consumer's reverse choice, and acting at a high cost to the society as a whole. Most of the related prior studies have focused on predicting review manipulation through data mining methods, and research from a consumer perspective is insufficient. However, since the possibility of manipulation of reviews perceived by consumers can affect the usefulness of reviews, it can provide important implications for online word-of-mouth management regardless of whether it is false or not. Therefore, in this study, we analyzed whether there is a difference between the review evaluated by the consumer as being manipulated and the general review, and verified whether the manipulated review negatively affects the review usefulness. For empirical analysis, 34,711 online book reviews on the LibraryThing website were analyzed using multilevel logistic regression analysis and Poisson regression analysis. As a result of the analysis, it was found that there were differences in product level, reviewer level, and review level factors between reviews that consumers perceived as being manipulated and reviews that were not. In addition, manipulated reviews have been shown to negatively affect review usefulness.

• Journal of KIISE
• /
• v.44 no.4
• /
• pp.339-343
• /
• 2017

As the number of customers who write fake reviews is increasing, online shopping sites have difficulty in providing reliable reviews. Fake reviews are called review spam, and they are written to promote or defame the product. They directly affect sales volume of the product; therefore, it is important to detect review spam. Review spam detection methods suggested in prior researches were only based on an international site even though review spam is a widespread problem in domestic shopping sites. In this paper, we have presented new review features of the domestic shopping site NAVER, and we have applied the formerly introduced method to this site for performing an evaluation.

• Knowledge Management Research
• /
• v.21 no.3
• /
• pp.177-195
• /
• 2020

The social consciousness on fake reviews has triggered researchers to suggest ways to cope with them by analyzing contents of fake reviews or finding ways to discover them by means of structural characteristics of them. This research tried to collect data from blog posts in Naver and detect habitual patterns users use unconsciously by variables extracted from blogs and blog posts by a machine learning model and wanted to use the technique in predicting fake reviews. Data analysis showed that there was a very high relationship between the number of all the posts registered in the blog of the writer of the related writing and the date when it was registered. And, it was found that, as model to detect advertising reviews, Random Forest is the most suitable. If a review is predicted to be an advertising one by the model suggested in this research, it is very likely that it is fake review, and that it violates the guidelines on investigation into markings and advertising regarding recommendation and guarantee in the Law of Marking and Advertising. The fact that, instead of using analysis of morphemes in contents of writings, this research adopts behavior analysis of the writer, and, based on such an approach, collects characteristic data of blogs and blog posts not by manual works, but by automated system, and discerns whether a certain writing is advertising or not is expected to have positive effects on improving efficiency and effectiveness in detecting fake reviews.

• Information Systems Review
• /
• v.23 no.4
• /
• pp.127-146
• /
• 2021

Fake news isexpanded and reproduced rapidly regardless of their authenticity by the characteristics of modern society, called the information age. Assuming that 1% of all news are fake news, the amount of economic costs is reported to about 30 trillion Korean won. This shows that the fake news isvery important social and economic issue. Therefore, this study aims to develop an automated detection model to quickly and accurately verify the authenticity of the news. To this end, this study crawled the news data whose authenticity is verified, and developed fake news prediction models using word embedding (Word2Vec, Fasttext) and deep learning algorithms (LSTM, BiLSTM). Experimental results show that the prediction model using BiLSTM with Word2Vec achieved the best accuracy of 84%.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2014.05a
• /
• pp.89-93
• /
• 2014

Reviews are very important for potential consumer' making choices. They are also used by manufacturers to find problems of their products and to collect competitors' business information. But someone write fake reviews to mislead readers to make wrong choices. Therefore detecting fake reviews is an important problem for the E-commerce sites. Support Vector Machines (SVMs) are very important text classification algorithms with excellent performance. In this paper, we propose a new incremental algorithm based on weight and the extension of Karush-Kuhn-Tucker(KKT) conditions and Convex Hull for online Review Spam Detection. Finally, we analyze its performance in theory.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.19 no.6
• /
• pp.1-8
• /
• 2019

Recently, consumer distrust has deepened due to blog posts focusing only on public relations due to 'viral marketing'. In addition, marketing projects such as false writing or exaggerated use of the latter phase are one of the most popular programs in 2016 as they are cheaper and more effective than newspaper and TV ads, and the size of advertising costs is set to be a major means of advertising at '3 trillion 394.1 billion won. From this 'viral marketing,' it has become an Internet environment that needs tools to filter information. The fake review filtering application ADDAVICHI presented in this paper extracts, analyzes, and presents blog keywords, total number of searches, reliability and satisfaction when users search for content such as "event" and "taste restaurant." Reliability shows the number of ad posts on a blog, the total number of posts, and satisfaction shows a clean post with confidence divided into positive and negative posts. Finally, the keyword shows a list of the top three words in the review from a positive post. In this way, it helps users interpret information away from advertising.