Back to results list
Show full item record
Please use this identifier to cite or link to this item:
|Title:||Dimensions of restaurant customer experience and emotions : an application of text analytics to fine-dining restaurant online reviews||Authors:||Oh, Mun Hyang||Degree:||Ph.D.||Issue Date:||2019||Abstract:||This study aims to (1) identify clusters in the semantic network of online reviews of fine-dining restaurants to reshape the dimensions of the restaurant experience, (2) determine basic emotions in online reviews of fine-dining restaurants and compare the performance of machine learning algorithms for text classification and (3) examine the semantic network for each emotion to understand the experiences involved in each emotion. Firstly, this study intends to determine the underlying dimensionality in online reviews regarding fine-dining restaurant experiences in Hong Kong. This study used 19,194 online reviews for data analysis and adopted semantic network analysis (SNA). Diverse and specific dimensions, such as ambiance, service, food, drinks, desserts, view, location, occasions, reputation and price, were explored. Secondly, this study identified the basic emotions in online reviews of Cantonese fine-dining restaurants and compared the performance of two machine learning algorithms for text classification. Emotions, such as joy, sadness, disgust, surprise and anger, appeared with 72% prediction accuracy. Emotions in fine-dining restaurants in Hong Kong were biased towards 'joy', indicating that obtaining hedonic value from food consumption experience could be the key motive for sharing fine-dining restaurant experiences. The comparison of the accuracy of the machine learning algorithms showed that support vector machine demonstrated better performance than the naive Bayes algorithm did. Thirdly, this study aims to investigate underlying stories within each emotion by adopting SNA. All five types of emotions in this study were correlated with service, food and reputation issues whether they were good or bad. In other words, people perceived service, food and reputation as the core aspects of a fine-dining restaurant experience. Location and private seats were related to joy, and good wine pairing was related to surprise. In addition, reviewers indicated their visit/not visit intention with joy/disgust.
The findings of this study provide four significant contributions to theory and practice. Firstly, the study offers a comprehensive framework of dimensionality for the fine-dining restaurant experience. This dimensionality is identified with a large volume of textual data. Secondly, the study extends the application of SNA to hospitality. Thirdly, the application of machine learning is one of the study's contributions to knowledge. Additionally, the findings can be useful for future studies that wish to adopt machine learning algorithms for text classification. Lastly, this study offers contributions that add to the dearth of existing literature on the application of cognitive appraisal theory to understand the behaviour that underlies the generation of electronic word-of-mouth (eWOM). Findings imply that subjective judgement and emotions motivate the generation of eWOM. This study has several implications for restaurant practitioners. In practice, although customers focus on certain similar fine-dining restaurant experiences, the frequently addressed aspects differ according to the type of ethnic restaurant. This finding indicates that restaurant practitioners should develop different strategies to adapt. Additionally, attentiveness and friendliness are important aspects of service quality. Most reviewers use these words to describe the service quality of fine-dining restaurants. Thus, practitioners can use this finding when educating their staff. Furthermore, the findings indicate that food has become increasingly specialised, and customers showcase this change. Restaurant managers can apply this finding to menu design by developing additional beverage and dessert menus and training their staff to improve knowledge on such menus. In terms of the dimension of the physical environment, people generally noticed comprehensive aspects, such as view, ambiance and decoration, instead of detailed features, such as lighting, music, temperature or interior design. Restaurant managers can arrange their interior on the basis of this observation to provide a holistic experience instead of focusing on one stimulus.
|Subjects:||Hong Kong Polytechnic University -- Dissertations
Text processing (Computer science)
|Pages:||xvi, 230 pages : color illustrations|
|Appears in Collections:||Thesis|
View full-text via https://theses.lib.polyu.edu.hk/handle/200/10211
Citations as of May 22, 2022
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.