RESUMO
Objective Tofacilitatetheradiologiststoinquireinformationrelatedtobreastdiseasesandimprovetheefficiencyand accuracyofsearchingdata.Methods Inthispaper,amultilevelstrategyhybridquestion-and-answermodelwasusedtodesignanautomatic question-and-answersysteminthefieldofbreastdiseasesimaging.Thefirstlayerofthismodelwasthequick matchingofquestion sentencesandFAQknowledgebase.Thesecondlayerwastoobtaintheanswersinthebasicknowledgebaseaccordingtothethresholdsetby processingthequestionsandcalculatingthesimilarityofthequestions.Thethirdlayerwastoacquiretheanswerparagraphfromthe webdocumentreturnedbythefull-textsearchengine.Results Thetestresultsshowedthattheaccuracyofthesystemansweracquisitionreached 85%.Thesystemcouldgivesatisfactoryanswerstotheproblemsexistingintheknowledgebase.Conclusion Theintelligentquestion answeringsystemforbreastdiseaseisconvenientandfast,anditisaneffectivetoolforradiologiststoinquiretheknowledgeofbreast diseases.