Automated Financial Analysis Using GPT-4

被引：1

作者：

Noels, Sander ^{[1
,2
]}

Merlevede, Adriaan ^{[2
]}

Fecheyr, Andrew ^{[2
]}

Vanhalst, Maarten ^{[2
]}

Meerlaen, Nick ^{[2
]}

Viaene, Sebastien ^{[2
]}

De Bie, Tijl ^{[1
]}

机构：

[1] Univ Ghent, Dept Elect & Informat Syst, B-9000 Ghent, Belgium

[2] Silverfin, Gaston Crommenlaan 12, B-9050 Ghent, Belgium

来源：

MACHINE LEARNING AND KNOWLEDGE DISCOVERY IN DATABASES: APPLIED DATA SCIENCE AND DEMO TRACK, ECML PKDD 2023, PT VII | 2023年 / 14175卷

关键词：

GPT-4; Automated Financial Analysis; Prompt Engineering;

D O I：

10.1007/978-3-031-43430-3_28

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

We introduce a novel application employing GPT-4 for automated financial analysis, aiming to improve the evaluation process of financial performance for businesses and accountants. Our application interprets financial ratios over a two-year period and generates actionable advice. Through specific design objectives and prompt engineering, we ensure the generated advice is concise, language-flexible, and informative. Integrated as a proof-of-concept (PoC) into Silverfin's accounting cloud service, our application showcases the potential of large language models in the specialized fields of finance and accounting.

引用

页码：345 / 349

页数：5