Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2506.23273 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909907354124288 |
|---|---|
| author | Nguyen, Quang Hung Trinh, Phuong Anh Mai, Phan Quoc Hung Trinh, Tuan Phong |
| author_facet | Nguyen, Quang Hung Trinh, Phuong Anh Mai, Phan Quoc Hung Trinh, Tuan Phong |
| contents | Despite the advancements of large language models, text2sql still faces many challenges, particularly with complex and domain-specific queries. In finance, database designs and financial reporting layouts vary widely between financial entities and countries, making text2sql even more challenging. We present FinStat2SQL, a lightweight text2sql pipeline enabling natural language queries over financial statements. Tailored to local standards like VAS, it combines large and small language models in a multi-agent setup for entity extraction, SQL generation, and self-correction. We build a domain-specific database and evaluate models on a synthetic QA dataset. A fine-tuned 7B model achieves 61.33\% accuracy with sub-4-second response times on consumer hardware, outperforming GPT-4o-mini. FinStat2SQL offers a scalable, cost-efficient solution for financial analysis, making AI-powered querying accessible to Vietnamese enterprises. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_23273 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | FinStat2SQL: A Text2SQL Pipeline for Financial Statement Analysis Nguyen, Quang Hung Trinh, Phuong Anh Mai, Phan Quoc Hung Trinh, Tuan Phong Artificial Intelligence Despite the advancements of large language models, text2sql still faces many challenges, particularly with complex and domain-specific queries. In finance, database designs and financial reporting layouts vary widely between financial entities and countries, making text2sql even more challenging. We present FinStat2SQL, a lightweight text2sql pipeline enabling natural language queries over financial statements. Tailored to local standards like VAS, it combines large and small language models in a multi-agent setup for entity extraction, SQL generation, and self-correction. We build a domain-specific database and evaluate models on a synthetic QA dataset. A fine-tuned 7B model achieves 61.33\% accuracy with sub-4-second response times on consumer hardware, outperforming GPT-4o-mini. FinStat2SQL offers a scalable, cost-efficient solution for financial analysis, making AI-powered querying accessible to Vietnamese enterprises. |
| title | FinStat2SQL: A Text2SQL Pipeline for Financial Statement Analysis |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2506.23273 |