請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/90023
標題: | 基於使用者生成內容進行餐廳選址排序 RESPECT: Restaurant Placement Ranking via Leveraging User-Generated Contents |
作者: | 江雨柔 Yu-Jo Chiang |
指導教授: | 魏志平 Chih-Ping Wei |
關鍵字: | 新店家選址,餐廳選址排序,社群分析,使用者生成內容,屬性層級情感分析,深度學習, New store location selection,Restaurant placement ranking,Community detection,User generated content,Aspect-based sentiment analysis,Deep Learning, |
出版年 : | 2023 |
學位: | 碩士 |
摘要: | 在餐飲業日益普及的背景下,如何在百家爭鳴中脫穎而出成為重要的商業議題,策略性地選址規劃變得格外關鍵。除了從供給與需求數據作為決策考量外,在社群媒體發展快速的年代,分析使用者生成的內容(UGC)來了解消費者在乎的面相以及對應的滿意度更是重要的資訊管道。透過屬性層級的情感分析(ABSA)技術的不斷改進,我們現在能夠更快且更真實地去了解用戶的反饋與需求。
在預測新店面適合的位置時,我們使用了多種特徵,其中包括描述候選位置以及餐廳的表現和競爭力的相關資訊。然而,傳統的競爭力計算方法卻常常依賴於商業資料、商店密度或評分等單一因素。為了克服這些限制,我們提出了一種新的方法,採用了ABSA技術來分析客戶的評論和反饋。此外,我們的競爭者分析採用了社群偵測技術,使得競爭者群組得以更為準確地被識別。這種方法與傳統的基於類別的競爭者分析相比,能夠提供更精確和深入的競爭環境分析。 經過一系列實驗,我們的研究結果顯示我們所提出的方法(稱為RESPECT)的有效性,並進一步解釋所得出的分析結果。我們所提出的RESPECT方法不僅能夠顯著改進餐廳選址決策的準確性和相關性,還展現了利用ABSA技術來制定更細緻、更數據驅動策略的潛力。總結,本論文的貢獻在於提出了一種全新且高效的方法,將情感分析和競爭者分析有機地結合,並採用ABSA技術來實現對餐廳選址的精細化評估。我們的研究將為餐廳業務和選址規劃提供寶貴的洞察,幫助餐廳業主和相關利益相關者做出更明智的決策,以在競爭激烈的行業中脫穎而出。 In the context of the increasingly popular catering industry, how to stand out in the intense competition has become essential; hence, selecting optimal locations for opening brick-and-mortar restaurant branches becomes even more indispensable. Besides relying on supply and demand data as predictive factors, with the proliferation and increasing popularity of social media platforms, user-generated contents (UGC) has increasingly become an important information source to understand consumers’ preferences and satisfaction. With the use of aspect-based sentiment analysis (ABSA) methods, we now have a faster and more authentic means of comprehending user feedbacks and demands. In predicting suitable locations for new restaurant establishment, we employ multiple features, descriptions of candidate locations, restaurant performance, and competitiveness-related information. However, conventional methods rely on limited data sources, such as commercial data, shop density, or ratings. To overcome these limitations, we propose a novel approach that analyzes customer reviews and feedbacks. We also utilize review with community detection to offer more precise and in-depth analysis of the competitor landscape. Our research results demonstrate the effectiveness of our proposed method (referred to as RESPECT). RESPECT not only significantly enhances the accuracy restaurant placement decisions but also showcases the potential of using ABSA techniques to formulate more refined and data-driven strategies. In conclusion, the study offers valuable insights for restaurant businesses and site planning, aiding restaurant owners and relevant stakeholders in making more informed decisions to excel in the fiercely competitive industry. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/90023 |
DOI: | 10.6342/NTU202303218 |
全文授權: | 未授權 |
顯示於系所單位: | 資訊管理學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-111-2.pdf 目前未授權公開取用 | 1.67 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。