{"id":"https://openalex.org/W4413969363","doi":"https://doi.org/10.3233/shti251383","title":"Evaluation of Chunking and Embedding Strategies for Local Document Retrieval Using an Open-Source LLM in a Hospital","display_name":"Evaluation of Chunking and Embedding Strategies for Local Document Retrieval Using an Open-Source LLM in a Hospital","publication_year":2025,"publication_date":"2025-09-03","ids":{"openalex":"https://openalex.org/W4413969363","doi":"https://doi.org/10.3233/shti251383","pmid":"https://pubmed.ncbi.nlm.nih.gov/40899531"},"language":"en","primary_location":{"id":"doi:10.3233/shti251383","is_oa":true,"landing_page_url":"https://doi.org/10.3233/shti251383","pdf_url":null,"source":{"id":"https://openalex.org/S4210179765","display_name":"Studies in health technology and informatics","issn_l":"0926-9630","issn":["0926-9630","1879-8365"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310318577","host_organization_name":"IOS Press","host_organization_lineage":["https://openalex.org/P4310318577"],"host_organization_lineage_names":["IOS Press"],"type":"book series"},"license":"cc-by-nc","license_id":"https://openalex.org/licenses/cc-by-nc","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Studies in Health Technology and Informatics","raw_type":"book-chapter"},"type":"article","indexed_in":["crossref","pubmed"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://doi.org/10.3233/shti251383","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5095373164","display_name":"Jan Bossenz","orcid":"https://orcid.org/0009-0001-7946-8559"},"institutions":[{"id":"https://openalex.org/I68956291","display_name":"Martin Luther University Halle-Wittenberg","ror":"https://ror.org/05gqaka33","country_code":"DE","type":"education","lineage":["https://openalex.org/I68956291"]}],"countries":["DE"],"is_corresponding":true,"raw_author_name":"Jan Bossenz","raw_affiliation_strings":["Junior Research Group (Bio-) Medical Data Science, Faculty of Medicine, Martin-Luther-University Halle-Wittenberg, Halle (Saale), Germany"],"raw_orcid":"https://orcid.org/0009-0001-7946-8559","affiliations":[{"raw_affiliation_string":"Junior Research Group (Bio-) Medical Data Science, Faculty of Medicine, Martin-Luther-University Halle-Wittenberg, Halle (Saale), Germany","institution_ids":["https://openalex.org/I68956291"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5119533523","display_name":"Carlo G\u00fcnzl","orcid":null},"institutions":[{"id":"https://openalex.org/I68956291","display_name":"Martin Luther University Halle-Wittenberg","ror":"https://ror.org/05gqaka33","country_code":"DE","type":"education","lineage":["https://openalex.org/I68956291"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Carlo G\u00fcnzl","raw_affiliation_strings":["Junior Research Group (Bio-) Medical Data Science, Faculty of Medicine, Martin-Luther-University Halle-Wittenberg, Halle (Saale), Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Junior Research Group (Bio-) Medical Data Science, Faculty of Medicine, Martin-Luther-University Halle-Wittenberg, Halle (Saale), Germany","institution_ids":["https://openalex.org/I68956291"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5051937514","display_name":"Fabian Berns","orcid":"https://orcid.org/0000-0002-7033-3789"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fabian Berns","raw_affiliation_strings":["medicalvalues GmbH, Karlsruhe, Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"medicalvalues GmbH, Karlsruhe, Germany","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5109013693","display_name":"Annemarie Bianka Weise","orcid":"https://orcid.org/0009-0003-6925-5941"},"institutions":[{"id":"https://openalex.org/I68956291","display_name":"Martin Luther University Halle-Wittenberg","ror":"https://ror.org/05gqaka33","country_code":"DE","type":"education","lineage":["https://openalex.org/I68956291"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Annemarie Weise","raw_affiliation_strings":["Junior Research Group (Bio-) Medical Data Science, Faculty of Medicine, Martin-Luther-University Halle-Wittenberg, Halle (Saale), Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Junior Research Group (Bio-) Medical Data Science, Faculty of Medicine, Martin-Luther-University Halle-Wittenberg, Halle (Saale), Germany","institution_ids":["https://openalex.org/I68956291"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5086531860","display_name":"Christian J\u00e4ger","orcid":"https://orcid.org/0000-0002-2740-8386"},"institutions":[{"id":"https://openalex.org/I68956291","display_name":"Martin Luther University Halle-Wittenberg","ror":"https://ror.org/05gqaka33","country_code":"DE","type":"education","lineage":["https://openalex.org/I68956291"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Christian J\u00e4ger","raw_affiliation_strings":["Junior Research Group (Bio-) Medical Data Science, Faculty of Medicine, Martin-Luther-University Halle-Wittenberg, Halle (Saale), Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Junior Research Group (Bio-) Medical Data Science, Faculty of Medicine, Martin-Luther-University Halle-Wittenberg, Halle (Saale), Germany","institution_ids":["https://openalex.org/I68956291"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5031585475","display_name":"Jan Kirchhoff","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jan Kirchhoff","raw_affiliation_strings":["medicalvalues GmbH, Karlsruhe, Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"medicalvalues GmbH, Karlsruhe, Germany","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5080416734","display_name":"Jan Christoph","orcid":"https://orcid.org/0000-0003-4369-3591"},"institutions":[{"id":"https://openalex.org/I68956291","display_name":"Martin Luther University Halle-Wittenberg","ror":"https://ror.org/05gqaka33","country_code":"DE","type":"education","lineage":["https://openalex.org/I68956291"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Jan Christoph","raw_affiliation_strings":["Junior Research Group (Bio-) Medical Data Science, Faculty of Medicine, Martin-Luther-University Halle-Wittenberg, Halle (Saale), Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Junior Research Group (Bio-) Medical Data Science, Faculty of Medicine, Martin-Luther-University Halle-Wittenberg, Halle (Saale), Germany","institution_ids":["https://openalex.org/I68956291"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5079157878","display_name":"Christoph Demus","orcid":"https://orcid.org/0000-0003-4319-3175"},"institutions":[{"id":"https://openalex.org/I68956291","display_name":"Martin Luther University Halle-Wittenberg","ror":"https://ror.org/05gqaka33","country_code":"DE","type":"education","lineage":["https://openalex.org/I68956291"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Christoph Demus","raw_affiliation_strings":["Junior Research Group (Bio-) Medical Data Science, Faculty of Medicine, Martin-Luther-University Halle-Wittenberg, Halle (Saale), Germany"],"raw_orcid":"https://orcid.org/0000-0003-4319-3175","affiliations":[{"raw_affiliation_string":"Junior Research Group (Bio-) Medical Data Science, Faculty of Medicine, Martin-Luther-University Halle-Wittenberg, Halle (Saale), Germany","institution_ids":["https://openalex.org/I68956291"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":8,"corresponding_author_ids":["https://openalex.org/A5095373164"],"corresponding_institution_ids":["https://openalex.org/I68956291"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.38618313,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"331","issue":null,"first_page":"91","last_page":"100"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12128","display_name":"AI in Service Interactions","score":0.4961000084877014,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T12128","display_name":"AI in Service Interactions","score":0.4961000084877014,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.14800000190734863,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11710","display_name":"Biomedical Text Mining and Ontologies","score":0.07410000264644623,"subfield":{"id":"https://openalex.org/subfields/1312","display_name":"Molecular Biology"},"field":{"id":"https://openalex.org/fields/13","display_name":"Biochemistry, Genetics and Molecular Biology"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/open-source","display_name":"Open source","score":0.8073709011077881},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6876119375228882},{"id":"https://openalex.org/keywords/embedding","display_name":"Embedding","score":0.6790415048599243},{"id":"https://openalex.org/keywords/chunking","display_name":"Chunking (psychology)","score":0.6392378211021423},{"id":"https://openalex.org/keywords/open-source-software","display_name":"Open source software","score":0.45578733086586},{"id":"https://openalex.org/keywords/information-retrieval","display_name":"Information retrieval","score":0.4457453489303589},{"id":"https://openalex.org/keywords/world-wide-web","display_name":"World Wide Web","score":0.33883580565452576},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.31817784905433655},{"id":"https://openalex.org/keywords/software","display_name":"Software","score":0.14947330951690674},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.1113176941871643}],"concepts":[{"id":"https://openalex.org/C3018397939","wikidata":"https://www.wikidata.org/wiki/Q3644502","display_name":"Open source","level":3,"score":0.8073709011077881},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6876119375228882},{"id":"https://openalex.org/C41608201","wikidata":"https://www.wikidata.org/wiki/Q980509","display_name":"Embedding","level":2,"score":0.6790415048599243},{"id":"https://openalex.org/C203357204","wikidata":"https://www.wikidata.org/wiki/Q1089605","display_name":"Chunking (psychology)","level":2,"score":0.6392378211021423},{"id":"https://openalex.org/C2988343187","wikidata":"https://www.wikidata.org/wiki/Q1130645","display_name":"Open source software","level":3,"score":0.45578733086586},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.4457453489303589},{"id":"https://openalex.org/C136764020","wikidata":"https://www.wikidata.org/wiki/Q466","display_name":"World Wide Web","level":1,"score":0.33883580565452576},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.31817784905433655},{"id":"https://openalex.org/C2777904410","wikidata":"https://www.wikidata.org/wiki/Q7397","display_name":"Software","level":2,"score":0.14947330951690674},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.1113176941871643}],"mesh":[{"descriptor_ui":"D009323","descriptor_name":"Natural Language Processing","qualifier_ui":null,"qualifier_name":null,"is_major_topic":true},{"descriptor_ui":"D009323","descriptor_name":"Natural Language Processing","qualifier_ui":null,"qualifier_name":null,"is_major_topic":true},{"descriptor_ui":"D009323","descriptor_name":"Natural Language Processing","qualifier_ui":null,"qualifier_name":null,"is_major_topic":true},{"descriptor_ui":"D016247","descriptor_name":"Information Storage and Retrieval","qualifier_ui":"Q000379","qualifier_name":"methods","is_major_topic":true},{"descriptor_ui":"D016247","descriptor_name":"Information Storage and Retrieval","qualifier_ui":"Q000379","qualifier_name":"methods","is_major_topic":true},{"descriptor_ui":"D016247","descriptor_name":"Information Storage and Retrieval","qualifier_ui":"Q000379","qualifier_name":"methods","is_major_topic":true},{"descriptor_ui":"D057286","descriptor_name":"Electronic Health Records","qualifier_ui":"Q000458","qualifier_name":"organization & administration","is_major_topic":true},{"descriptor_ui":"D057286","descriptor_name":"Electronic Health Records","qualifier_ui":"Q000458","qualifier_name":"organization & administration","is_major_topic":true},{"descriptor_ui":"D057286","descriptor_name":"Electronic Health Records","qualifier_ui":"Q000458","qualifier_name":"organization & administration","is_major_topic":true}],"locations_count":2,"locations":[{"id":"doi:10.3233/shti251383","is_oa":true,"landing_page_url":"https://doi.org/10.3233/shti251383","pdf_url":null,"source":{"id":"https://openalex.org/S4210179765","display_name":"Studies in health technology and informatics","issn_l":"0926-9630","issn":["0926-9630","1879-8365"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310318577","host_organization_name":"IOS Press","host_organization_lineage":["https://openalex.org/P4310318577"],"host_organization_lineage_names":["IOS Press"],"type":"book series"},"license":"cc-by-nc","license_id":"https://openalex.org/licenses/cc-by-nc","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Studies in Health Technology and Informatics","raw_type":"book-chapter"},{"id":"pmid:40899531","is_oa":false,"landing_page_url":"https://pubmed.ncbi.nlm.nih.gov/40899531","pdf_url":null,"source":{"id":"https://openalex.org/S4306525036","display_name":"PubMed","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I1299303238","host_organization_name":"National Institutes of Health","host_organization_lineage":["https://openalex.org/I1299303238"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Studies in health technology and informatics","raw_type":null}],"best_oa_location":{"id":"doi:10.3233/shti251383","is_oa":true,"landing_page_url":"https://doi.org/10.3233/shti251383","pdf_url":null,"source":{"id":"https://openalex.org/S4210179765","display_name":"Studies in health technology and informatics","issn_l":"0926-9630","issn":["0926-9630","1879-8365"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310318577","host_organization_name":"IOS Press","host_organization_lineage":["https://openalex.org/P4310318577"],"host_organization_lineage_names":["IOS Press"],"type":"book series"},"license":"cc-by-nc","license_id":"https://openalex.org/licenses/cc-by-nc","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Studies in Health Technology and Informatics","raw_type":"book-chapter"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4376877853","https://openalex.org/W1493891899","https://openalex.org/W4250928611","https://openalex.org/W166480398","https://openalex.org/W1612808768","https://openalex.org/W167327709","https://openalex.org/W1977393088","https://openalex.org/W4387839566","https://openalex.org/W4210922265","https://openalex.org/W2288962794"],"abstract_inverted_index":{"INTRODUCTION:":[0],"The":[1,45,182],"integration":[2],"of":[3,60,83,256],"Retrieval-Augmented":[4,62],"Generation":[5,63],"(RAG)":[6,64],"into":[7],"domain-specific":[8,225],"systems":[9],"enables":[10],"context-aware":[11],"and":[12,20,42,88,97,107,116,125,131,149,158,212,218,231,240,258],"traceable":[13],"information":[14],"retrieval.":[15],"This":[16],"study":[17,50],"explores":[18],"chunking":[19,257],"embedding":[21,100,219],"strategies":[22],"for":[23,56,77,174,250],"a":[24,61,245],"RAG-based":[25],"question-answering":[26],"system":[27,67,249],"tailored":[28],"to":[29,70,73,166],"administrative":[30,251],"documents":[31,85],"at":[32],"University":[33],"Hospital":[34],"Halle,":[35],"focusing":[36],"on":[37],"model":[38],"selection,":[39],"parameter":[40,123,167,259],"tuning,":[41],"retrieval":[43,172,223],"performance.":[44],"insights":[46],"gained":[47],"from":[48],"this":[49],"should":[51],"serve":[52],"as":[53],"the":[54,57,138,203],"foundation":[55],"future":[58,254],"development":[59],"based":[65],"chatbot":[66],"that":[68,190],"aims":[69],"facilitate":[71],"access":[72],"document":[74],"pool":[75],"contents":[76],"hospital":[78],"staff.":[79],"METHODS:":[80],"A":[81],"corpus":[82],"1,219":[84],"was":[86],"preprocessed":[87],"chunked":[89],"using":[90,104,128],"varied":[91],"parameters,":[92],"including":[93],"soft/hard":[94],"character":[95],"limits":[96],"overlaps.":[98],"Eight":[99],"models":[101,114],"were":[102,118,233],"evaluated":[103],"Similarity":[105],"Score":[106],"Maximum":[108],"Marginal":[109],"Relevance":[110],"(MMR)":[111],"retrievers.":[112],"Top":[113],"Jinaai-v3":[115,152,191,230],"Aari1995":[117,136,232],"further":[119],"analyzed":[120],"across":[121,146],"eight":[122],"configurations":[124],"ensemble":[126,214],"retrievers":[127,170],"weight":[129],"(w)":[130],"context":[132],"(c)":[133],"parameters.":[134],"RESULTS:":[135],"reached":[137],"highest":[139],"Top10":[140],"score":[141,205],"(92.3%)":[142],"with":[143,163],"stable":[144],"performance":[145],"chunk":[147],"sizes":[148],"retriever":[150,206],"configurations.":[151,215],"showed":[153,184],"slightly":[154],"stronger":[155],"Top5":[156],"(84.6%)":[157],"Top3":[159],"(76.9%)":[160],"scores":[161],"but":[162],"greater":[164],"sensitivity":[165],"variations.":[168],"Ensemble":[169],"improved":[171],"quality":[173],"both":[175,210,229],"models,":[176],"particularly":[177],"when":[178],"tuned":[179],"via":[180],"w-values.":[181],"c-parameter":[183],"negligible":[185],"influence.":[186],"Runtime":[187],"evaluation":[188],"revealed":[189],"generated":[192],"vector":[193],"stores":[194],"more":[195],"than":[196,200],"four":[197],"times":[198],"faster":[199],"Aari1995.":[201],"Overall,":[202],"similarity":[204],"consistently":[207],"outperformed":[208],"MMR,":[209],"standalone":[211],"in":[213,224,237],"CONCLUSION:":[216],"Chunking":[217],"choices":[220],"significantly":[221],"affect":[222],"RAG":[226,248],"systems.":[227],"While":[228],"effective,":[234],"they":[235],"differed":[236],"stability,":[238],"accuracy,":[239],"efficiency.":[241],"Findings":[242],"support":[243],"deploying":[244],"locally":[246],"executable":[247],"use,":[252],"guiding":[253],"optimization":[255],"robustness.":[260]},"counts_by_year":[],"updated_date":"2026-04-29T09:16:38.111599","created_date":"2025-10-10T00:00:00"}
