five

Data set of pro-policies in four places

收藏
科学数据银行2025-03-26 更新2026-04-23 收录
下载链接:
https://www.scidb.cn/detail?dataSetId=fb9aefc30d0f4a31bbe58770da22fc76
下载链接
链接失效反馈
官方服务:
资源简介:
[Purpose] The existing policy analysis methods rely heavily on manual annotation and alignment comparison, resulting in low efficiency and easy errors. This study aims to improve the efficiency of policy information retrieval, achieve intelligent analysis and comparison of policies, and provide precise decision support for policy formulation by constructing a structured policy knowledge base. [Method] This study takes the policy of benefiting enterprises as an example and proposes a framework based on the big language model for efficient comparison of relevant policies. The framework includes the following steps: 1) Building a knowledge base; 2) Retrieval and storage; 3) Answer generation. [Results] Through the validation of the national, Beijing, Shanghai, and Shenzhen policy datasets for benefiting enterprises, the framework proposed in this paper automatically integrates multiple policies and can analyze policy semantics to construct a database, helping to complete policy matching and analysis. The Chroma RAG model used in this study demonstrates significant advantages in Hit@1 The target has reached 60% Hit@3 The indicator reached 76%, and the MRR indicator reached 71.13%. In the comparison of retrieval methods, the performance of our research model is superior to traditional models such as Tf idf, Word2Vec, USE, BERT, SBERT, DPR, SimCSE, etc., highlighting the superiority of our method. By Hit@1 The Hit@3 Compared with evaluation indicators such as MRR, the framework and retrieval method proposed in this study have significant advantages. [Limitations] The research is mainly based on cross-sectional data and cannot fully reflect the dynamic changes in the policy implementation process, which limits the in-depth analysis of policy effects. [Conclusion] The construction of knowledge base and policy comparison based on big language model can effectively improve the intelligent analysis and comparison effect of policy texts, especially in the construction of policy knowledge base and policy comparison support, providing significant decision support for policy makers.
提供机构:
duan yong kang
创建时间:
2025-03-26
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作