A Hybrid RAG System Integrating Knowledge Graph and Vector Retrieval: Based on Solution Technical Documents

A Hybrid RAG System Integrating Knowledge Graph and Vector Retrieval: Based on Solution Technical Documents | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article A Hybrid RAG System Integrating Knowledge Graph and Vector Retrieval: Based on Solution Technical Documents Cheonsu Jeong This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9476748/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract As enterprise environments increasingly manage large volumes of technical documentation across multiple heterogeneous sources, users face growing challenges in efficiently locating relevant information. This study proposes and implements a hybrid Retrieval-Augmented Generation (RAG) system that integrates a knowledge graph with vector retrieval to enhance information access in solution technical documents. The proposed system combines a Neo4j-based knowledge graph with a ChromaDB-based vector database, enabling both structural relationship exploration and semantic similarity search. A domain-specific knowledge graph is constructed by extracting entities and relationships from technical documents, including cross-document references derived from HTML anchor links. In parallel, vector embeddings are generated using a multilingual embedding model to support semantic retrieval across diverse document types. The system employs a hybrid ranking mechanism based on Reciprocal Rank Fusion (RRF) to effectively integrate graph-based and vector-based retrieval results. A seven-stage query processing pipeline is designed to analyze user queries, perform graph traversal and vector search, fuse results, and generate natural language responses using a large language model. Experimental evaluation on a multi-source technical document dataset demonstrates that the proposed hybrid approach significantly improves retrieval performance compared to single-method baselines, particularly in complex queries requiring cross-document reasoning. The results indicate that integrating knowledge graph and vector retrieval within a hybrid RAG framework provides an effective solution for enhancing information retrieval in enterprise technical documentation environments. Artificial Intelligence and Machine Learning Hybrid RAG Knowledge Graph Vector Retrieval Information Retrieval Reciprocal Rank Fusion (RRF) Full Text Additional Declarations The authors declare no competing interests. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9476748","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":626608457,"identity":"8f2bc003-85b7-4a94-9132-353fafc940a3","order_by":0,"name":"Cheonsu Jeong","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA2klEQVRIiWNgGAWjYJCCAwwMFnIScC4PcVokjEnTAgQSiTOI1mLOfvbh4YIaifSZ7b0PH/xgsJNn4Dn7AK8Wy550g8MzjknkzuY5bmzYw5Bs2MDbboBXi8GBNIbDPGwSufMk0tgkeBiYExj42fA7zOD8M6CWfxLpchJp7D//MNQToeUG0BbeNokEaaAtzDwMhxMYeNvwa7GcAbRlZp+E4cyeY8zSMgbHDdt4juHXYs6fxvy54JuNvMTxNsaPbyqq5fl50gg4DIiZUbgEfIKhZRSMglEwCkYBFgAA3Lw5hFWA61MAAAAASUVORK5CYII=","orcid":"","institution":"SAMSUNG SDS","correspondingAuthor":true,"prefix":"","firstName":"Cheonsu","middleName":"","lastName":"Jeong","suffix":""}],"badges":[],"createdAt":"2026-04-20 22:38:28","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-9476748/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9476748/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107531449,"identity":"e94627d4-a138-4b15-91f5-d09ba80903b3","added_by":"auto","created_at":"2026-04-22 10:27:41","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":905661,"visible":true,"origin":"","legend":"","description":"","filename":"GraphRAGHybridSearch.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9476748/v1_covered_dcd67479-aa74-4fd8-8101-561c9dd3e941.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eA Hybrid RAG System Integrating Knowledge Graph and Vector Retrieval: Based on Solution Technical Documents\u003c/strong\u003e\u003c/p\u003e","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"SAMSUNG SDS","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Hybrid RAG, Knowledge Graph, Vector Retrieval, Information Retrieval, Reciprocal Rank Fusion (RRF)","lastPublishedDoi":"10.21203/rs.3.rs-9476748/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9476748/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAs enterprise environments increasingly manage large volumes of technical documentation across multiple heterogeneous sources, users face growing challenges in efficiently locating relevant information. This study proposes and implements a hybrid Retrieval-Augmented Generation (RAG) system that integrates a knowledge graph with vector retrieval to enhance information access in solution technical documents. The proposed system combines a Neo4j-based knowledge graph with a ChromaDB-based vector database, enabling both structural relationship exploration and semantic similarity search. A domain-specific knowledge graph is constructed by extracting entities and relationships from technical documents, including cross-document references derived from HTML anchor links. In parallel, vector embeddings are generated using a multilingual embedding model to support semantic retrieval across diverse document types. The system employs a hybrid ranking mechanism based on Reciprocal Rank Fusion (RRF) to effectively integrate graph-based and vector-based retrieval results. A seven-stage query processing pipeline is designed to analyze user queries, perform graph traversal and vector search, fuse results, and generate natural language responses using a large language model. Experimental evaluation on a multi-source technical document dataset demonstrates that the proposed hybrid approach significantly improves retrieval performance compared to single-method baselines, particularly in complex queries requiring cross-document reasoning. The results indicate that integrating knowledge graph and vector retrieval within a hybrid RAG framework provides an effective solution for enhancing information retrieval in enterprise technical documentation environments.\u003c/p\u003e","manuscriptTitle":"A Hybrid RAG System Integrating Knowledge Graph and Vector Retrieval: Based on Solution Technical Documents","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-22 10:27:05","doi":"10.21203/rs.3.rs-9476748/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"a519f406-0f41-4a39-a91a-8051cf3b1bcc","owner":[],"postedDate":"April 22nd, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":66805508,"name":"Artificial Intelligence and Machine Learning"}],"tags":[],"updatedAt":"2026-04-22T10:27:05+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-22 10:27:05","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9476748","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9476748","identity":"rs-9476748","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}