• Asia Marketing Journal
• /
• v.15 no.4
• /
• pp.61-101
• /
• 2014

This study investigates when and how disagreements in online customer ratings prompt more favorable product evaluations. Among the three metrics of volume, valence, and variance that feature in the research on online customer ratings, volume and valence have exhibited consistently positive patterns in their effects on product sales or evaluations (e.g., Dellarocas, Zhang, and Awad 2007; Liu 2006). Ratings variance, or the degree of disagreement among reviewers, however, has shown rather mixed results, with some studies reporting positive effects on product sales (e.g., Clement, Proppe, and Rott 2007) while others finding negative effects on product evaluations (e.g., Zhu and Zhang 2010). This study aims to resolve these contradictory findings by introducing preference heterogeneity as a possible moderator and causal attribution as a mediator to account for the moderating effect. The main proposition of this study is that when preference heterogeneity is perceived as high, a disagreement in ratings is attributed more to reviewers' different preferences than to unreliable product quality, which in turn prompts better quality evaluations of a product. Because disagreements mostly result from differences in reviewers' tastes or the low reliability of a product's quality (Mizerski 1982; Sen and Lerman 2007), a greater level of attribution to reviewer tastes can mitigate the negative effect of disagreement on product evaluations. Specifically, if consumers infer that reviewers' heterogeneous preferences result in subjectively different experiences and thereby highly diverse ratings, they would not disregard the overall quality of a product. However, if consumers infer that reviewers' preferences are quite homogeneous and thus the low reliability of the product quality contributes to such disagreements, they would discount the overall product quality. Therefore, consumers would respond more favorably to disagreements in ratings when preference heterogeneity is perceived as high rather than low. This study furthermore extends this prediction to the various levels of average ratings. The heuristicsystematic processing model so far indicates that the engagement in effortful systematic processing occurs only when sufficient motivation is present (Hann et al. 2007; Maheswaran and Chaiken 1991; Martin and Davies 1998). One of the key factors affecting this motivation is the aspiration level of the decision maker. Only under conditions that meet or exceed his aspiration level does he tend to engage in systematic processing (Patzelt and Shepherd 2008; Stephanous and Sage 1987). Therefore, systematic causal attribution processing regarding ratings variance is likely more activated when the average rating is high enough to meet the aspiration level than when it is too low to meet it. Considering that the interaction between ratings variance and preference heterogeneity occurs through the mediation of causal attribution, this greater activation of causal attribution in high versus low average ratings would lead to more pronounced interaction between ratings variance and preference heterogeneity in high versus low average ratings. Overall, this study proposes that the interaction between ratings variance and preference heterogeneity is more pronounced when the average rating is high as compared to when it is low. Two laboratory studies lend support to these predictions. Study 1 reveals that participants exposed to a high-preference heterogeneity book title (i.e., a novel) attributed disagreement in ratings more to reviewers' tastes, and thereby more favorably evaluated books with such ratings, compared to those exposed to a low-preference heterogeneity title (i.e., an English listening practice book). Study 2 then extended these findings to the various levels of average ratings and found that this greater preference for disagreement options under high preference heterogeneity is more pronounced when the average rating is high compared to when it is low. This study makes an important theoretical contribution to the online customer ratings literature by showing that preference heterogeneity serves as a key moderator of the effect of ratings variance on product evaluations and that causal attribution acts as a mediator of this moderation effect. A more comprehensive picture of the interplay among ratings variance, preference heterogeneity, and average ratings is also provided by revealing that the interaction between ratings variance and preference heterogeneity varies as a function of the average rating. In addition, this work provides some significant managerial implications for marketers in terms of how they manage word of mouth. Because a lack of consensus creates some uncertainty and anxiety over the given information, consumers experience a psychological burden regarding their choice of a product when ratings show disagreement. The results of this study offer a way to address this problem. By explicitly clarifying that there are many more differences in tastes among reviewers than expected, marketers can allow consumers to speculate that differing tastes of reviewers rather than an uncertain or poor product quality contribute to such conflicts in ratings. Thus, when fierce disagreements are observed in the WOM arena, marketers are advised to communicate to consumers that diverse, rather than uniform, tastes govern reviews and evaluations of products.

• Journal of Intelligence and Information Systems
• /
• v.26 no.4
• /
• pp.1-25
• /
• 2020

Sentiment Analysis (SA) is a Natural Language Processing (NLP) task that analyzes the sentiments consumers or the public feel about an arbitrary object from written texts. Furthermore, Aspect-Based Sentiment Analysis (ABSA) is a fine-grained analysis of the sentiments towards each aspect of an object. Since having a more practical value in terms of business, ABSA is drawing attention from both academic and industrial organizations. When there is a review that says "The restaurant is expensive but the food is really fantastic", for example, the general SA evaluates the overall sentiment towards the 'restaurant' as 'positive', while ABSA identifies the restaurant's aspect 'price' as 'negative' and 'food' aspect as 'positive'. Thus, ABSA enables a more specific and effective marketing strategy. In order to perform ABSA, it is necessary to identify what are the aspect terms or aspect categories included in the text, and judge the sentiments towards them. Accordingly, there exist four main areas in ABSA; aspect term extraction, aspect category detection, Aspect Term Sentiment Classification (ATSC), and Aspect Category Sentiment Classification (ACSC). It is usually conducted by extracting aspect terms and then performing ATSC to analyze sentiments for the given aspect terms, or by extracting aspect categories and then performing ACSC to analyze sentiments for the given aspect category. Here, an aspect category is expressed in one or more aspect terms, or indirectly inferred by other words. In the preceding example sentence, 'price' and 'food' are both aspect categories, and the aspect category 'food' is expressed by the aspect term 'food' included in the review. If the review sentence includes 'pasta', 'steak', or 'grilled chicken special', these can all be aspect terms for the aspect category 'food'. As such, an aspect category referred to by one or more specific aspect terms is called an explicit aspect. On the other hand, the aspect category like 'price', which does not have any specific aspect terms but can be indirectly guessed with an emotional word 'expensive,' is called an implicit aspect. So far, the 'aspect category' has been used to avoid confusion about 'aspect term'. From now on, we will consider 'aspect category' and 'aspect' as the same concept and use the word 'aspect' more for convenience. And one thing to note is that ATSC analyzes the sentiment towards given aspect terms, so it deals only with explicit aspects, and ACSC treats not only explicit aspects but also implicit aspects. This study seeks to find answers to the following issues ignored in the previous studies when applying the BERT pre-trained language model to ACSC and derives superior ACSC models. First, is it more effective to reflect the output vector of tokens for aspect categories than to use only the final output vector of [CLS] token as a classification vector? Second, is there any performance difference between QA (Question Answering) and NLI (Natural Language Inference) types in the sentence-pair configuration of input data? Third, is there any performance difference according to the order of sentence including aspect category in the QA or NLI type sentence-pair configuration of input data? To achieve these research objectives, we implemented 12 ACSC models and conducted experiments on 4 English benchmark datasets. As a result, ACSC models that provide performance beyond the existing studies without expanding the training dataset were derived. In addition, it was found that it is more effective to reflect the output vector of the aspect category token than to use only the output vector for the [CLS] token as a classification vector. It was also found that QA type input generally provides better performance than NLI, and the order of the sentence with the aspect category in QA type is irrelevant with performance. There may be some differences depending on the characteristics of the dataset, but when using NLI type sentence-pair input, placing the sentence containing the aspect category second seems to provide better performance. The new methodology for designing the ACSC model used in this study could be similarly applied to other studies such as ATSC.

• The Journal of Art Theory & Practice
• /
• no.2
• /
• pp.27-48
• /
• 2004

Robert Rauschenberg's artistic career has often been regarded as having reached its culmination when the artist won the first prize at the 1964 Venice Biennale. With this victory, Rauschenberg triumphantly entered the pantheon of all-American artists and firmly secured his position in the history of American art. On the other hand, despite the artist's ongoing new experiments in his art, the seemingly precocious ripeness in his career has led the critical discourses on Rauschenberg's art to the artist's early works, most of which were done in the mid-1950s and the 1960s. The crux of Rauschenberg criticism lies not only in focusing on the artist's 50's and 60's works, but also in its large dismissal of the significance of the imagery that the artist employed in his works. As art historians Roger Cranshaw and Adrian Lewis point out, the critical discourse of Rauschenberg either focuses on the formalist concerns on the picture plane, or relies on the "culturalist" interpretation of Rauschenberg's imagery which emphasizes the artist's "Americanness." Recently, a group of art historians centered around October has applied Charles Sanders Peirce's semiotics as art historical methodology and illuminated the indexical aspects of Rauschenberg's work. The semantic inquiry into Rauschenberg's imagery has also been launched by some art historians who seek the clues in the artist's personal context. The first half of this essay will examine the previous criticism on Rauschenberg's art and the other half will discuss the artist's 1955 work Rebus, which I think intersects various critical concerns of Rauschenberg's work, and yet defies the closure of discourses in one direction. The categories of signs in the semiotics of Charles Sanders Peirce and the discourse of Jean-Francois Lyotard will be used in discussing the meanings of Rebus, not to search for the semantic readings of the work, hut to make an analogy in terms of the paradoxical structures of both the work and the theory. The definitions of rebus is as follows: Rebus 1. a representation or words or syllables by pictures of object or by symbols whose names resemble the intended words or syllables in sound; also: a riddle made up wholly or in part of such pictures or symbols. 2. a badge that suggests the name of the person to whom it belongs. Webster's Third New International Dictionary of the English Language Unabridged. Since its creation in 1955, Robert Rauschenberg's Rebus has been one of the most intriguing works in the artist's oeuvre. This monumental 'combine' painting($6feet{\times}10feet$ 10.5 inches) consists of three panels covered with fabric, paper, newspaper, and printed reproductions. On top of these, oil paints, pencil and crayon drawings connect each section into a whole. The layout of the images is overall horizontal. Starting from a torn election poster, which is partially read as "THAT REPRE," on the far left side of the painting. Rebus leads us to proceed from the left to the right, the typical direction of reading in a Western context. Along with its seemingly proper title. Rebus, the painting has triggered many art historians to seek some semantic readings of it. These art historians painstakingly reconstruct the iconography based on the artist's interviews, (auto)biography, and artistic context of his works. The interpretation of Rebus varies from a 'image-by-image' collation with a word to a more general commentary on Rauschenberg's work overall, such as a work that "bridges between art and life." Despite the title's allusion to the legitimate purpose of the painting as a decoding of the imagery into sound, Rebus, I argue, actually hinders a reading of it. By reading through Peirce to Rauschenberg, I will delve into the subtle anxiety between words and images in their works. And on this basis, I suggest Rauschenberg's strategy in playing Rebus is to hide the meaning of the imagery rather than to disclose it.

• Journal of Intelligence and Information Systems
• /
• v.25 no.1
• /
• pp.109-125
• /
• 2019

As interest on intelligent search engines increases, various studies have been conducted to extract and utilize the features related to products intelligencely. In particular, when users search for goods in e-commerce search engines, the 'color' of a product is an important feature that describes the product. Therefore, it is necessary to deal with the synonyms of color terms in order to produce accurate results to user's color-related queries. Previous studies have suggested dictionary-based approach to process synonyms for color features. However, the dictionary-based approach has a limitation that it cannot handle unregistered color-related terms in user queries. In order to overcome the limitation of the conventional methods, this research proposes a model which extracts RGB values from an internet search engine in real time, and outputs similar color names based on designated color information. At first, a color term dictionary was constructed which includes color names and R, G, B values of each color from Korean color standard digital palette program and the Wikipedia color list for the basic color search. The dictionary has been made more robust by adding 138 color names converted from English color names to foreign words in Korean, and with corresponding RGB values. Therefore, the fininal color dictionary includes a total of 671 color names and corresponding RGB values. The method proposed in this research starts by searching for a specific color which a user searched for. Then, the presence of the searched color in the built-in color dictionary is checked. If there exists the color in the dictionary, the RGB values of the color in the dictioanry are used as reference values of the retrieved color. If the searched color does not exist in the dictionary, the top-5 Google image search results of the searched color are crawled and average RGB values are extracted in certain middle area of each image. To extract the RGB values in images, a variety of different ways was attempted since there are limits to simply obtain the average of the RGB values of the center area of images. As a result, clustering RGB values in image's certain area and making average value of the cluster with the highest density as the reference values showed the best performance. Based on the reference RGB values of the searched color, the RGB values of all the colors in the color dictionary constructed aforetime are compared. Then a color list is created with colors within the range of ${\pm}50$ for each R value, G value, and B value. Finally, using the Euclidean distance between the above results and the reference RGB values of the searched color, the color with the highest similarity from up to five colors becomes the final outcome. In order to evaluate the usefulness of the proposed method, we performed an experiment. In the experiment, 300 color names and corresponding color RGB values by the questionnaires were obtained. They are used to compare the RGB values obtained from four different methods including the proposed method. The average euclidean distance of CIE-Lab using our method was about 13.85, which showed a relatively low distance compared to 3088 for the case using synonym dictionary only and 30.38 for the case using the dictionary with Korean synonym website WordNet. The case which didn't use clustering method of the proposed method showed 13.88 of average euclidean distance, which implies the DBSCAN clustering of the proposed method can reduce the Euclidean distance. This research suggests a new color synonym processing method based on RGB values that combines the dictionary method with the real time synonym processing method for new color names. This method enables to get rid of the limit of the dictionary-based approach which is a conventional synonym processing method. This research can contribute to improve the intelligence of e-commerce search systems especially on the color searching feature.