Skip navigation

DSpace JSPUI

DSpace preserves and enables easy and open access to all types of digital content including text, images, moving images, mpegs and data sets

Learn More
DSpace logo
English
中文
  • Browse
    • Communities
      & Collections
    • Publication Year
    • Author
    • Title
    • Subject
  • Search TDR
  • Rights Q&A
    • My Page
    • Receive email
      updates
    • Edit Profile
  1. NTU Theses and Dissertations Repository
  2. 管理學院
  3. 資訊管理學系
Please use this identifier to cite or link to this item: http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/88548
Title: 加密貨幣網紅社群貼文主題變化模型
Topic Evolution for Cryptocurrency Influencers on Social Media Platforms
Authors: 潘躍升
Yueh-Sheng Pan
Advisor: 李瑞庭
Anthony J. T. Lee
Keyword: 主題變化,加密貨幣,區塊鏈,時間主題模型,圖注意力網路,記憶機制,
topic evolution,cryptocurrency,blockchain,temporal topic model,graph attention network,memory mechanism,
Publication Year : 2023
Degree: 碩士
Abstract: 近年來,區塊鏈受到廣泛的關注,尤其是加密貨幣在區塊鏈生態系統中扮演著重要的腳色。創辦人、項目所有者、愛好者或名人等網紅常透過社交媒體平台(如 Twitter),公開分享他們對加密貨幣相關事件和市場趨勢的觀點,提供了許多寶貴的資訊。然而,大多數的推文長度十分簡短且包含縮寫,要從推文中有效地提取有價值的見解具有相當的挑戰性。一些既有方法並非針對短文本進行設計,或部分既有方法未能充分利用資料的特性來進行主題追蹤。為了解決上述的這些問題,我們提出了一個創新的方法,名為動態主題記憶網絡,自動從推文中提取主題的變化並進行時間主題分析。模型的框架包含四個階段,分別是詞嵌入、圖卷積、鍵值記憶網絡和時間主題分析,在無監督的方式下透過匯集詞嵌入語意關係的相關資訊,並通過關鍵組件和核心向量來生成一致性和多樣性的主題。實驗結果顯示,我們的方法在主題連貫性、主題多樣性、主題質量和質化分析上均優於現有的方法。我們的方法可以幫助區塊鏈架構師和加密貨幣交易者掌握最新動態,並可協助加密貨幣市場經理或去中心化金融經理發覺新興主題和趨勢,以擬定其行銷與定位策略。
Blockchain has received considerable attention in recent years, especially with cryptocurrency playing a crucial role in the blockchain ecosystem. Influencers, including founders, project owners, enthusiasts, and celebrities, publicly share their opinions on crypto-related events and market trends through social media platforms such as Twitter, providing valuable information as a reference in these ever-changing crypto events. Since most tweets are brief and include abbreviations, capturing and extracting valuable insights from such tweet posts can be challenging. Moreover, some previous methods are not designed for short text, while some may not fully leverage the characteristics of text data. To resolve such problems, in this study, we propose a novel approach, called Dynamic Topic Memory Network (DTMN), to automatically extract the evolution of topics derived from tweet posts and perform temporal topic analysis. The proposed framework contains four phases, namely, token embeddings, graph convolution, key-value memory network, and temporal topic analysis to aggregate relevant information from neighboring token dependencies and construct coherence and diverse topics extracted by key components and core vectors in an unsupervised manner. The experimental results show that our proposed approach outperforms the state-of-the-art methods in terms of topic coherence, topic diversity, topic quality, and qualitative evaluation. Our approach can help blockchain architects and cryptocurrency traders stay updated on the latest developments, and the crypto marketing managers or decentralized finance managers identify emerging topics and trends for implementing their marketing and positioning strategies.
URI: http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/88548
DOI: 10.6342/NTU202302045
Fulltext Rights: 同意授權(限校園內公開)
metadata.dc.date.embargo-lift: 2028-07-25
Appears in Collections:資訊管理學系

Files in This Item:
File SizeFormat 
ntu-111-2.pdf
  Restricted Access
834.47 kBAdobe PDFView/Open
Show full item record


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

社群連結
聯絡資訊
10617臺北市大安區羅斯福路四段1號
No.1 Sec.4, Roosevelt Rd., Taipei, Taiwan, R.O.C. 106
Tel: (02)33662353
Email: ntuetds@ntu.edu.tw
意見箱
相關連結
館藏目錄
國內圖書館整合查詢 MetaCat
臺大學術典藏 NTU Scholars
臺大圖書館數位典藏館
本站聲明
© NTU Library All Rights Reserved