被動投資是追蹤市場指數所涵蓋的股票或債券為投資標的，通常以ETF建構投資組合，不預測市場，秉持著資產配置、分散風險與長線持有的投資觀念。由於市場無法長期精準的被預測，想要尋找市場上被低估的個股，來賺取超額報酬，將會越來越難達成。且多數國家對於金融法規是高度監管，應用於投資領域的聊天機器人的研究相對較少，其中又以主動投資的研究佔大多數，故本研究以被動投資理財聊天機器人為主題，解決客戶於QA互動時的專業問題。
　　本研究以中文建構被動投資的知識本體論，用於儲存領域知識，採用個案金融機構的投資理念，以及財富管理專家的經驗與知識，來追蹤對話中的重要實體(Entity)。建構2種不同技術的聊天機器人，分別為RASA NLU與ChatGPT Prompt Engineering，我們提出了投資理財領域的聊天機器人設計方法，實現協助投資人理財規劃與回答理財知識，從客戶經常詢問理財顧問的問題中，比較傳統自然語言理解(NLU)與新穎大型語言模型(LLM)的表現。
　　我們使用混淆矩陣(Confusion Matrix)分析聊天機器人意圖分類的精確率(Precision)與召回率(Recall)，以及實體提取的準確率(Accuracy)。並分析聊天機器人模型對於意圖分類的信心程度(Confidence)，最後以質性分析對話流程的使用性(Usability)。研究發現，LLM聊天機器人在意圖分類與實體提取的表現較好，LLM意圖分類的召回率(Recall)高於RASA有20%，LLM實體提取的準確率(Accuracy)高於RASA有12%。對話流程方面，RASA表現較結構化、LLM表現較多樣化。

　　Passive investment typically involves constructing investment portfolios with ETFs, without predicting the market. It adheres to the investment principles of asset allocation, risk diversification, and long-term holding. As there is relatively limited research on chatbots applied to the investment field, and the majority of existing studies focus on active investment, this research centers on a passive investment financial chatbot. The aim is to address clients' professional inquiries during QA interactions.
　　This research constructs a ontology for passive investment in Chinese, utilizing the investment principles of a specific financial institution, as well as the expertise and knowledge of wealth management specialists. Two different chatbot technologies were developed, namely RASA NLU and ChatGPT Prompt Engineering, to assist investors in financial planning and provide answers to financial knowledge. By comparing the performance of traditional Natural Language Understanding (NLU) and novel Large Language Models (LLM) based on common inquiries from clients directed to financial advisors, the study aims to achieve these objectives.
　　We used a Confusion Matrix to analyze the Precision and Recall of intent classification and the Accuracy of entity extraction for the chatbot. We also assessed the Confidence level of the chatbot model in intent classification and conducted a qualitative analysis of Usability in the dialog. The study revealed that the LLM chatbot performed better in both intent classification and entity extraction compared to RASA. The Recall for intent classification was 20% higher in LLM, and the Accuracy of entity extraction was 12% higher in LLM when compared to RASA. In terms of dialog, RASA exhibited a more structured approach, while LLM demonstrated greater diversity.

論文審定書 i
致謝 ii
摘要 iii
Abstract iv
目錄 v
圖次 vii
表次 ix
第一章　緒論 1
第一節　研究背景 1
第二節　研究動機 3
第三節　研究目的 4
第二章　文獻探討 5
第一節　聊天機器人在投資理財領域的應用 5
第二節　聊天機器人 6
第三節　知識本體論 8
第三章　研究方法 11
第一節　被動投資知識本體論定義 11
第二節　意圖與實體設計 20
第三節　實驗評估方法 22
第四章　RASA聊天機器人設計 23
第一節　系統架構 23
第二節　意圖訓練 23
第三節　實體提取 33
第四節　結構化查詢 35
第五節　自然語言處理 37
第六節　對話流程 39
第五章　LLM聊天機器人設計 45
第一節　提示指南 46
第二節　提示工程 46
第三節　對話流程 49
第六章　聊天機器人評估 53
第一節　意圖分類分析 53
第二節　實體提取分析 58
第三節　對話流程比較 59
第七章　結論 61
第一節　研究貢獻 61
第二節　未來展望 62
第八章 參考文獻 63

Al-Zubaide, H., & Issa, A. A. (2011). Ontbot: Ontology based chatbot. International Symposium on Innovations in Information and Communications Technology,
Alberts, L. K. (1995). YMIR: An ontology for engineering design.
Altinok, D. (2018). An ontology-based dialogue management system for banking and finance dialogue systems. arXiv preprint arXiv:1804.04838.
Chuang, W. T., & Yang, J. (2000). Extracting sentence segments for text summarization: a machine learning approach. Proceedings of the 23rd annual international ACM SIGIR conference on Research and development in information retrieval,
Cruz, C., & Nicolle, C. (2008). XML-IS: Ontology-Based Integration Architecture. WEBIST (1),
Ellis, C. D. (2016). The Index Revolution. Wiley.
Fayyad, U., Piatetsky-Shapiro, G., & Smyth, P. (1996). From data mining to knowledge discovery in databases. AI magazine, 17(3), 37-37.
Feldman, R., & Sanger, J. (2007). The text mining handbook: advanced approaches in analyzing unstructured data. Cambridge university press.
Gao, Z., & Liang, Y. (2015). The ontology construction approach for the Chinese Tax Knowledge Domain. 2015 12th International Conference on Fuzzy Systems and Knowledge Discovery (FSKD),
Genesereth, M. R., & Nilsson, N. J. (2012). Logical foundations of artificial intelligence. Morgan Kaufmann.
Gruber, T. R. (1993). A translation approach to portable ontology specifications. Knowledge acquisition, 5(2), 199-220.
Guan, A., Wang, Y., & Yang, L. (2015). Automatic term extraction for chinese opera domain ontology. 2015 12th International Conference on Fuzzy Systems and Knowledge Discovery (FSKD),
Gupta, A., & Sharma, D. (2019). Customers’ Attitude towards Chatbots in Banking Industry of India. International Journal of Innovative technology and exploring Engineering, 8(11), 1222-1225.
Hien, L. T., Ly, L. T. T., Pham-Nguyen, C., Le Dinh, T., & Gia, H. T. (2020). Towards chatbot-based interactive what-and how-question answering systems: the adobot approach. 2020 RIVF International Conference on Computing and Communication Technologies (RIVF),
Hwerbi, K. (2020). Ontology-Based Chatbot for Disaster Management: Use Case CoronaVirus. arXiv preprint arXiv:2011.02340.
Joshi, A., Kunchukuttan, A., Bhattacharyya, P., & Carman, M. J. (2015). SarcasmBot: An open-source sarcasm-generation module for chatbots. WISDOM Workshop at KDD,
Kao, A., & Poteet, S. R. (2007). Natural language processing and text mining. Springer Science & Business Media.
KPMG. (2023). Pulse of Fintech H2'22 (Global analysis of fintech investment, Issue. https://assets.kpmg.com/content/dam/kpmg/tw/pdf/2023/02/pulse-of-fintech-h2-22-web-file-en.pdf
L'Abbate, M., Thiel, U., & Kamps, T. (2005). Can proactive behavior turn chatterbots into conversational agents? IEEE/WIC/ACM International Conference on Intelligent Agent Technology,
Nazir, A., Khan, M. Y., Ahmed, T., Jami, S. I., & Wasi, S. (2019). A novel approach for ontology-driven information retrieving chatbot for fashion brands. Int. J. Adv. Comput. Sci. Appl. IJACSA, 10(9).
Nirenburg, S., & Raskin, V. (2001). Ontological semantics, formal ontology, and ambiguity. Proceedings of the international conference on Formal Ontology in Information Systems-Volume 2001,
Noy, N. F., & McGuinness, D. L. (2001). Ontology development 101: A guide to creating your first ontology. In: Stanford knowledge systems laboratory technical report KSL-01-05 and ….
Patidar, M., Agarwal, P., Vig, L., & Shroff, G. (2017). Correcting Linguistic Training Bias in an FAQ-bot using LSTM-VAE. DMNLP Workshop of ECML-PKDD,
Peters, M. E., Neumann, M., Iyyer, M., Gardner, M., Clark, C., Lee, K., & Zettlemoyer, L. (2018). Deep contextualized word representations. arXiv preprint arXiv:1802.05365.
Quamar, A., Lei, C., Miller, D., Ozcan, F., Kreulen, J., Moore, R. J., & Efthymiou, V. (2020). An ontology-based conversation system for knowledge bases. Proceedings of the 2020 ACM SIGMOD International Conference on Management of Data,
Quamar, A., Özcan, F., Miller, D., Moore, R. J., Niehus, R., & Kreulen, J. (2020). Conversational BI: an ontology-driven conversation system for business intelligence applications. Proceedings of the VLDB Endowment, 13(12), 3369-3381.
Reshmi, S., & Balakrishnan, K. (2016). Implementation of an inquisitive chatbot for database supported knowledge bases. sādhanā, 41, 1173-1178.
Saha, G. K. (2007). Web ontology language (OWL) and semantic web. Ubiquity, 2007(September), 1-1.
Sarbabidya, S., & Saha, T. (2020). Role of chatbot in customer service: a study from the perspectives of the banking industry of Bangladesh. International review of business research papers, 16(1), 231-248.
Vysniauskas, E., & Nemuraite, L. (2006). Transforming ontology representation from OWL to relational database. Information technology and control, 35(3).
Wen, M.-H. (2018). A conversational user interface for supporting individual and group decision-making in stock investment activities. 2018 IEEE International Conference on Applied System Invention (ICASI),
Wu, Y., Wu, W., Xing, C., Zhou, M., & Li, Z. (2016). Sequential matching network: A new architecture for multi-turn response selection in retrieval-based chatbots. arXiv preprint arXiv:1612.01627.
Zhang, W., & Hamori, S. (2021). Crude oil market and stock markets during the COVID-19 pandemic: Evidence from the US, Japan, and Germany. International Review of Financial Analysis, 74, 101702.
Zhuge, H., Xing, Y., & Shi, P. (2008). Resource space model, OWL and database: Mapping and integration. ACM Transactions on Internet Technology (TOIT), 8(4), 1-31.