尚玉梅



摘要:為了提高圖書館管理系統(tǒng)在使用過程中的個(gè)性化和智能推薦需求,通過分析大數(shù)據(jù)挖掘?qū)嵤┓桨福x擇合適的數(shù)據(jù)挖掘工具,對圖書信息進(jìn)行預(yù)處理,并結(jié)合支持向量機(jī)和神經(jīng)網(wǎng)絡(luò)算法建立了個(gè)性化圖書信息推薦服務(wù)方案,然后進(jìn)行系統(tǒng)功能需求分析和系統(tǒng)整體架構(gòu)分析,最后進(jìn)行了實(shí)例展示和分析。研究發(fā)現(xiàn):采用改進(jìn)的SVM算法來實(shí)現(xiàn)圖書館的個(gè)性化數(shù)據(jù)挖掘,支持向量機(jī)算法在使用過程中具有監(jiān)督的、可擴(kuò)展和非線性的高效特性,能夠?qū)崿F(xiàn)非線性的多核心數(shù)據(jù)聚類效果,從而提高數(shù)據(jù)挖掘的學(xué)習(xí)能力;利用BP神經(jīng)網(wǎng)絡(luò)對處理后的數(shù)據(jù)樣本進(jìn)行適應(yīng)性訓(xùn)練,用戶在使用過程中給予一定的正向反饋,該決策分析體系根據(jù)反饋結(jié)果進(jìn)行不斷的自主學(xué)習(xí)并更新和優(yōu)化樣品數(shù)據(jù),實(shí)現(xiàn)了一個(gè)閉合的良性循環(huán);通過對設(shè)計(jì)的個(gè)性化圖書館推薦服務(wù)系統(tǒng)使用體驗(yàn)調(diào)查發(fā)現(xiàn):選A的讀者占比為58%,選B的讀者占35%,說明在使用過程中對于該個(gè)性化推薦系統(tǒng)滿意度超過了90%,能夠?yàn)樽x者用戶提供一定的借閱便利。
關(guān)鍵詞:圖書館;大數(shù)據(jù)挖掘;決策分析;個(gè)性化
中圖分類號:TP315
文獻(xiàn)標(biāo)志碼:A
ResearchonConstructionofEfficientLibraryBigDataMiningand
DecisionAnalysisSystemBasedonPersonalizedService
SHANGYumei
(
Library,ShanxiVocationalandTechnicalCollege,Xian710038,China
)
Abstract:Inordertoimprovetheuseoflibrarymanagementsystemintheprocessofpersonalizedandintelligentrecommendedrequirements,thisarticle,throughtheanalysisoflargedataminingplan,selectstheappropriatedataminingtoolstopreprocessthebooksinformation,andcombinessupportvectormachineandneuralnetworkalgorithmtoestablishpersonalizedbookinformationrecommendationserviceplan.Thenthesystemfunctiondemandanalysisandoverallsystemarchitectureanalysisarecompleted.Finallytheinstanceanalysisiscarriedoutanddisplayed.ItisfoundthattheimprovedSVMalgorithmcanbeusedtorealizepersonalizeddatamininginthelibrary,andtheSVMalgorithmhasthecharacteristicsofsupervised,extensibleandnonlinearefficiencyintheprocessofuse,andcanachievethenonlinearmulticoredataclusteringeffect,soastoimprovethelearningabilityofdatamining.Thedecisionanalysissystemcancontinuouslyindependentlylearn,updateandoptimizethesampledataaccordingtothefeedbackresults.Thusitachievesaclosedvirtuouscircle.Throughtheinvestigationontheuserexperienceofthedesignedpersonalizedlibraryrecommendationservicesystem,itisfoundthattheproportionofreaderswhochooseAis58%,andthatofreaderswhochooseBis35%,indicatingthatthesatisfactionofthepersonalizedrecommendationsystemexceeds90%,whichcanprovidecertainborrowingconvenienceforreadersandusers.
Keywords:library;bigdatamining;decisionanalysis;personalization
0引言
移動(dòng)互聯(lián)網(wǎng)和人工智能的發(fā)展對于信息的甄別效率有的新的要求,現(xiàn)代社會(huì)迫切需要在海量信息沖擊下如何在短時(shí)間內(nèi)獲取自己感興趣或者想要的信息[1]。近年來,智慧校園概念的提出便是智能化推薦和個(gè)性化定制需求的一個(gè)實(shí)踐方向。……