基於圖神經網路的細粒度論點探勘及其在預測力評估之應用

Abstract

在日常生活中，人們常常會討論對未來的展望或趨勢，如職業規劃、財務規劃或時尚趨勢，不僅如此，在各種專業領域中也很常需要對未來前景的推估，包括制定商業策略、探討社會變遷、經濟政策和環境保護等等。這些未來導向的分析在各領域的決策中發揮著至關重要的作用，然而，要評估這類前瞻性論述的品質是相當困難的。

在這些不同領域中，我們特別選擇一個具專業性的前瞻性分析文件：證券研究報告。這些報告由專業分析師撰寫，提供對上市公司的評估，分析其當前狀況，討論影響其表現的關鍵因素，並且最重要的是，對公司的未來前景和股票估值做出預測。鑑於這些報告在投資決策中的重要性，評估分析師意見的預測能力至關重要。然而，由於證券研究報告本身已具有較高的品質，如何進一步評估這些報告中論點的預測能力變得更加具有挑戰性。

本研究提出了一種結合細粒度論點探勘及情感分析的證券研究報告之分析方法。鑑於傳統方法在捕捉未來導向分析的細微差別存在侷限性，我們提出了一種細緻的論點單元分類方法,將其劃分為主張、前提和情景,並結合了獨特的情感分析框架。此外，我們還引入了時間維度,對市場事件的預期影響持續時間進行分類。為了進行這項研究，我們也建立了證券研究報告的標記資料集(Equity-AMSA)。本研究探討了在特定領域中資料標記能夠達到的細緻化程度以及在現今大型語言模型時代下細粒度標記資料的必要性，也探討我們提出的框架是否能較傳統方法在評估論點品質的任務中取得更好的表現。

實驗結果顯示了細粒度人工標記的重要性，尤其是針對情景的識別和情感分析而言。總結來說，我們提出的標記框架和資料集有助於更透徹地理解證券研究報告的內容及提升模型在預測力評估之表現。

In daily life, people often discuss future outlooks or trends, such as career planning, financial planning, or fashion trends. Moreover, in various professional fields, it is common to need to estimate future prospects, including formulating business strategies, exploring social changes, economic policies, and environmental protection. These forward-looking analyses play a crucial role in decision-making across different domains. However, evaluating the quality of such prospective discussions is quite challenging.

Among these different fields, we specifically choose a professional forward-looking analysis document: the equity research report. Authored by financial experts, these reports provide in-depth assessments of publicly traded companies, analyzing their current position, identifying key factors influencing their performance, and most importantly, offering predictions about their future growth and stock valuations. Given the critical role these reports play in investment decisions, it becomes crucial to evaluate the forecasting skill of the analysts' opinions. However, since equity research reports already possess a high level of quality, further assessing the forecasting skill of the arguments within these reports becomes even more challenging.

This paper introduces a novel approach to the analysis of equity research reports by integrating argument mining with sentiment analysis. Recognizing the limitations of traditional models in capturing the nuances of future-oriented analysis, we propose a refined categorization of argument units into claims, premises, and scenarios, coupled with a unique sentiment analysis framework. Furthermore, we incorporate a temporal dimension to categorize the anticipated impact duration of market events. To facilitate this study, we present the Equity Argument Mining and Sentiment Analysis (Equity-AMSA) dataset. Our research investigates the extent to which detailed domain-specific annotations can be provided, the necessity of fine-grained human annotations in the era of large language models, and whether our proposed framework can improve performance in downstream tasks over traditional methods.

Experimental results reveal the significance of manual annotations, especially for scenario identification and sentiment analysis. The study concludes that our annotation scheme and dataset contribute to a deeper understanding of forward-looking statements in equity research reports.