[Objective] Based on the research results related to the subject cross-database retrieval service, optimize the Lingo3G algorithm combined with the solr scoring rules to realize the cross-database knowledge integration and knowledge fingerprint service of the institutional repository. [Method] Analyze the actual needs of users, build a functional framework for knowledge integration analysis and visualization, select key technologies and methods to build a platform, and explore the feasibility of knowledge integration. [Results] Research and calculate the characteristics of knowledge fingerprints in the institutional knowledge base, organize and present knowledge fingerprints in a visual form, and realize cross-database knowledge integration services with third-party databases through clustering. [Limitations] Because the database structure and design methods of cross-database retrieval are different and many databases do not have public resource retrieval API interfaces, the limitations of cross-database retrieval cannot be generally solved. [Conclusion] Based on the Lingo3G clustering algorithm, it can provide a set of functional systems for the institutional knowledge base based on cross-database retrieval to effectively improve the integration of knowledge resources and mine knowledge service capabilities.