Household financial advice generation using retrieval-augmented generation

Miwa, Juriin; Li, Da; Kumamoto, Tadahiko; Kawai, Yukiko

Permalink

https://hdl.handle.net/2142/126223 Copy

Description

Title

Household financial advice generation using retrieval-augmented generation

Author(s)

• Miwa, Juriin
• Li, Da
• Kumamoto, Tadahiko
• Kawai, Yukiko

Issue Date

2025-03-11

Keyword(s)

• Financial Consultation
• Retrieval-Augmented Generation
• Advice Generation

Abstract

This study aims to enhance the quality of automated financial advice generation based on large language models (LLMs) by constructing a Retrieval-Augmented Generation (RAG) system from expert-written texts on household income and expenditure data. The construction of the RAG involves scraping financial planning (FP) consultation articles from the web, extracting not only the consultation text but also basic information such as the gender and family composition of the consulter, household account book images, and the advice provided by the FP. Next, income and expenditure data are extracted from the household account book images, and vector data are generated from Pinecone based on the consultation content and income and expenditure data. More accurate financial advice is generated by inputting the expanded consultation content into the LLM through the constructed vector database. In this paper, the implementation and verification of the proposed method for generating financial advice using RAG on OpenAI are discussed, exploring the applicability of RAG-based financial management support using text and tables.

Publisher

iSchools

Series/Report Name or Number

iConference 2025 Proceedings

Type of Resource

Other

Genre of Resource

Conference Poster

Language

eng

Handle URL

https://hdl.handle.net/2142/126223

Copyright and License Information

Copyright 2025 is held by Juriin Miwa, Da Li, Tadahiko Kumamoto, and Yukiko Kawai. Copyright permissions, when appropriate, must be obtained directly from the author.

Owning Collections