National Kaohsiung First University of Science and Technology
FirstTech Institutional RepositoryNot a member yet
18584 research outputs found
Sort by
A Study of Smart Retail Using Facial Image Analysis
No full text[[abstract]]有鑑於行政院商業4.0發展策略的執行，說明我們正迎接著商業新時代的來臨，其中智慧零售的應用在此更顯為重要，透過串聯線上線下的商務模式、整合實體通路與電子商務的服務，都將成為往後發展的重點。並期望將多元通路整合為全通路的營運模式，結合數據分析建立消費者行為模式模型，打造零距離的互動體驗，藉以提升服務價值。本研究搜集了包含性別、年齡標籤的公開人臉資料庫，分別進行性別及年齡模型的訓練，並利用CaffeFramework及GoogLeNetNetwork進行實作；此外結合Kinect攝影機的骨架資訊來進行人臉位置的框選及動作辨識的體感操作體驗；最後結合南仁湖企業哪購電子商務的歷史訂單資料，並依據自定義的客戶群，進行商品間的關聯規則尋找，來做為商品推薦的依據。在本研究的實驗中，我們將實作的性別及年齡模型與目前的公開模型（如LAPmodel等）進行測試與比較，實驗結果顯示其在公開人臉資料庫（如imdb-wiki500k+等）可達86.12%及76.39%的準確率，平均執行時間為1.04秒，雖然並未優於目前的公開模型，但考量到即時辨識的應用需求上，其更適合本系統應用。在動作辨識體感操作的部分，使用LV3D特徵結合DTW的方法，在跨受測者的任意開始與結束，都能有90%以上的辨識成功率。最後在歷史訂單關聯分析上，我們以支持度及信賴度作為評估條件，進行商品推薦的實作，並期許能為企業在行銷上帶來更多的助力。[[abstract]]InviewoftheimplementationoftheExecutiveYuanBusiness4.0strategy,wearefacingtheadventofaneweraofbusiness,inwhichtheapplicationofsmartretailingismoreimportanthere.Andhopetointegratethemulti-channelpathtoOmni-Channel,combinedwithdataanalysistoestablishconsumerbehaviormodelinordertoenhanceservicevalue.Inthisstudy,wecollectedthepublicopenfacedatabasethatcontaininggenderandagelabel.WeusedCaffeFrameworkandGoogLeNetNetworktotrainingthegenderandagemodels.Inaddition,wecombinedtheskeletontrackingtolocatefacepositionandmotionrecognitionviaKinectsensor.Finally,weusedthehistoricalorderdatatorunassociationrulebetweentheproductsthatsupportbyNargoe-commerce.Thenimplementtheproductrecommendationbasedonacustomizedcustomergroups.Intheexperiment,wetestedandcomparedthegenderandagemodelwithpublicopenmodel(suchasLAPmodel).Theresultsshowthatinthepublicopenfacedatabase(suchasimdb-wiki500k+etc.)reach86.12%and76.39%accuracyrate.Theaverageexecutiontimeis1.04seconds.AlthoughitisnotbetterthantheLAPmodel,butitismoresuitableforreal-timeapplicationofthissystem.Inpartofthemotionrecognition,weusedLV3DframelevelfeatureandDTWmethod.Theresultsshowthatinthecross-testedandunknownstartandendmotioncanreachupto90%accuracyrate.Finally,inthehistoricalorderdataassociationruleanalysis,weusedthesupportvaluesandconfidencevaluestoimplementationofproductrecommendation.Andwehopetobringmorehelpfortheenterpriseinthemarketing
The exploration and application of Blockchain
No full text[[abstract]]近年來，隨著金融科技的蓬勃發展，使用互聯網金融的人口年年攀升，然而，傳統金融產業正面臨新科技(FinTech)的破壞式創新的衝擊，新的金融科技時代來臨。在這一波的金融科技的革新技術－區塊鏈出現。帶動台灣金融業對區塊鏈的重視，區塊鏈衍然成為金融科技業的關鍵技術。所以，本研究的目的是利用區塊鏈技術，以「分散式帳本」、「密碼學機制」、「智能合約」與「共識機制」四大構成要素為基礎歸納出以下六個特性「不可篡改性」、「不可否認性」、「可追蹤性」、「私密性」、「正確性」及「完整性」。採方法論以紮根理論為基礎，透過現有已使用區塊鏈機制做個案實例分析，最後產出一套區塊鏈應用程序發展概念的標準流程，讓未來藉助這個方法論來給國內各行各業在應用區塊鏈的時候有一個遵循的法則。本研究發現未來各產業想要導入區塊鏈，可以透過本研究流程圖為一個參考的依循，未來各個產業都可以參考這個流程而導入區塊鏈的應用。關鍵字：紮根理論、區塊鏈、信任機制、案例分析[[abstract]]Inrecentyears,withthevigorousdevelopmentoffinancialtechnology,theuseofInternetfinancehasriseneveryyear.However,thetraditionalfinancialindustryisfacingtheimpactofthenewtechnologicalinnovation(FinTech),theneweraoffinancialtechnology.Inthiswaveoffinancialtechnologyinnovationtechnologyblockchainappears.LedtheTaiwanfinancialindustrytotheblockchainofattention,blockchainhasbecomethekeytechnologyoffinancialtechnologyindustry.Therefore,thepurposeofthisstudyistousetheblockchaintechnologyto"DistidutedLedgers"、"Cryptography"、"SmartContracts"and"Consensus"basedonthefourelementsofthefollowingsixcharacteristics"non-tampering"、"non-repudiation"、"traceability"、"privacy"、"correctness"and"completeness".Basedonthetheoryofrooting,themethodhasbeenusedtoanalyzethecasechain,andfinallyproduceasetofstandarddevelopmentprocessoftheconceptofblockchainapplicationdevelopment,sothatthefutureuseofthismethodtothedomesticindustryinallindustriesWhentheblockchainhasaruletofollow.Thisstudyfoundthatthefutureindustrywanttoimporttheblockchain,throughtheflowchartofthisstudyasareferencetofollowthefutureofeachindustrycanrefertothisprocessandtheintroductionoftheapplicationoffastchain.Keywords:GroundedTheory、Blockchain、TrustMechanism、CaseAnalysi
The Impact of Innovations on B2B Customer Relationship
No full text[[abstract]]在企業對企業市場中，客戶維繫一直以來都是關係行銷的探討核心。近年來，越來越多企業開始重視與客戶維持長期互惠關係。供應商為了創造更多利潤，有效的經營與維持忠實顧客是勢在必行的。與客戶建立長期互惠關係往往都被視為成功的關鍵，因此，企業具備卓越的關係管理策略是十分重要的。客戶維繫與商業應用在學術研究中並沒有廣泛地探討。本研究將分析關係價值、信任、承諾與客戶維繫之因果關係。本研究以台灣紡織業為研究對象，共回收383份有效問卷，目的是為了瞭解供應商具備技術與服務之創新，對於強化客戶的維繫有無顯著影響。而實證結果發現，(1)客戶之去留與商業定位有關，(2)供應商越能滿足客戶需求，留住客戶的可能性越大，綜合實證結果分析，客戶維繫的關鍵是客戶與供應商在創新上秉持相同的理念。[[abstract]]Inbusinesstobusinessmarkets,customerretentionhasbeenthecenterofrelationshipmarketing.Duringthepastfewyears,thereisanincreasingnumberofbusinessesshiftingtheiremphasisonnurturinglong-lastingrelationshipswiththeirbusinesscustomers.Inthebeliefofmakingmoreprofits,itisessentialforsupplierstomanageandmaintainloyalB2Bcustomers.Therearemanymarketersregardingthatthefundamentalofbusinesssuccessisbuildingalong-termrelationshipwithbusinesscustomers.Applyingastrategicrelationshipmanagementisimportanttobuyersandsellerstowardsbusinessindustries.However,wediscoveredthatthedevelopmentofcustomerretentionanditspotentialapplicationtothebusinesstobusinesssettingshasnotbeenwidelyexplored.Thepurposeofthisstudyistofilltheresearchgapbyempiricallyanalyzingtherelativitybetweenrelationshipvalueandothergenerallystudiedvariablesoftrust,commitmentandcustomerretention.Inthisstudy,respondentswerecompanieswithtotalrecoveryof383validquestionnaires.ThequestionnairewhichwastargetingTaiwan’stextileindustryisattemptingtorealizebusinesssupplierswithadvancedinnovationshavesignificantorinsignificanteffecttoreinforcebusinesscustomers’retentiontothesuppliersinthebusinesstobusinesscontext.Thefindingsfromthedataanalysissuggestthat(1)theretentionofbusinesscustomersdependsontheirbusinessorientationofcompanies.(2)Themoretoreachcustomers’requirements,themorepossibilitytoretain.Toconcludethisstudy,thekeytoreachcustomerretentionisthatthecustomersmustsharethesameperspectiveoninnovationswiththeirsuppliers
Empirical Analysis of Aggregative Technical Indicators in Taiwanese Stock Market
No full text[[abstract]]本論文認為順勢而為的交易策略、紀律操作，為散戶股市投資獲利最重要的方式，因此本研究利用綜合9種技術指標方法，如移動平均線、KD指標、MACD指標等作為進場點的條件，並搭配紀律性的出場條件，從台灣股票市場1990年1月1日至2016年12月31日之間的每個交易日、每檔個股中，挑選符合條件的個股進行投資，並逐步篩選出最適的綜合技術指標，並找出較為穩定報酬和較多投資機會及比較穩定精準的投資策略。本研究發現在多頭訊號且20MA上彎、5MA、10MA與20MA上彎、近5日(近一周)總成交量是前6-10日(上周)成交量的2.8倍、進場日長紅K棒實體大於3%、KD黃金交叉且KD均小於80、MACD紅柱時進場，且在20MA下彎時出場可獲得較高的報酬率。[[abstract]]Momentumtradingstrategieswithdisciplinedrulesisthekeyfactortoearnprofits.Asaresults,thispaperemploys9technicalindicators,suchasmovingaverages(MA),StochasticOscillator,MovingAverageConvergence/Divergence(MACD)toconstructthebuysignalsaswellasthesellsignals.Using150stockstradedinTaiwanesestockmarketfromJanuary1,1990toDecember31,2016,thispaperconstructsthebuyandsellsignalsaccordingto9technicalindicatorsandbullishpatternandempiricallytesttheirperformance.Theempiricalresultsshowthatusingthetradingstrategiesbasedontechnicalindicatorsandbullishpatterncouldprovidebetterperformance
The Commercial Value of Circular Economy:An Example of the Recycling of Wasted PU Foam
No full text[[abstract]]本研究探討回收廢泡棉在循環經濟中的商業價值並以技術可行之再生產品為例進行經濟效益分析。研究重點著重於（1）資源回收與循環經濟是一個重要的發展趨勢，（2）找出經濟價值才能解決資源回收的問題，（3）再生產品的經濟效益與現行處理方式效益進行比較分析。結果顯示以2016年環保署公告的廢機動車輛與回收電冰箱統計資料估算之廢泡棉數量約4,898噸計算，焚化處理費用高達新台幣2億4仟餘萬元，若以本研究中最具經濟效益之再生產品估算，經濟效益將可達新台幣3億3仟餘萬元，因此將廢泡棉重新導入市場內，不但能解決現行焚化處理產生之問題，更能有效創造循環經濟之價值。[[abstract]]ThestudyexploresthecommercialvalueofwastedPUfoamrecyclinginthecirculareconomy,andanalyzesitseconomicbenefitsbytakingthetechnicallyfeasiblerenewableproductasanexample.Theresearchfocuseson(1)Resourcerecyclingandcirculareconomyareimportantdevelopmenttrends.(2)Findouttheeconomyvalueandthenbeabletosolveproblemsfromrecycling.(3)Comparativeanalysisbetweeneconomicbenefitsofrenewableproductsandbenefitsoncurrenttreatmentmethod.ThestatisticaldataannouncedbyEPAin2016showthewastedPUfoamfromthescrapedvehiclesandrecycledfridgesareabout4,989tonnesandthecostforincineratingthesewastedPUfoamreachesNTD330millioninaccordancewithmystudy.Therefore,tore-enterthewastedPUfoamintothemarketisnotonlysolveproblemsarisedfromcurrentincinerationtreatmentbutcreatethecirculareconomicvalue
Factors Analysis of the Risk of Accounts Receivable: An Example of the Foreign Semiconductor Company
No full text[[abstract]]本研究主要分析台灣中小企業的外商半導體公司的應收帳款風險的影響因子，並應用卡方檢定與二元羅吉斯迴歸分析與觀察期間為2013至2016年。經由實證研究結果顯示，在卡方檢定中達到具有統計上的顯著，包括：客戶公司資本額、客戶公司員工數量、交易年份以及交易月份，同時於二元羅吉斯迴歸結果中，發現顯著的變數分別有客戶所在區域、客戶公司員工數量、付期45天與付期75天，其中正向關係的變數為客戶所在區域以及客戶公司員工數量，存在負向關係的是付期45天與付期75天。[[abstract]]Thisempiricalintroducesamodeloftheriskofaccountsreceivableusingchi-squaretestandbinarylogisticregression.Theresultsindicatethatthesevariablesincludedinthechi-squaretestmodel,amountofcapital,numberofemployees,tradingyearandmonthweresignificantlydifferent.Inbinarylogisticregressionmodel,thelocatedinAsia,numberofemployees,45-day’spayperiodand75-day’spayperiodweresignificantlydifferentwhichshowedthattheseresultsalsoprovidethattheriskofaccountsreceivablearepositivelyassociatedwiththelocatedinAsiaandnumberofemployeesandnegativelyrelatedwith45-day’spayperiodand75-day’spayperiod
An Analysis on Financial Indicator Performance of Life Insurance Company in Taiwan: Holding and Non-Holding Companies
No full text[[abstract]]本文分析2011年到2015年24家台灣地區壽險公司的經營績效對公司每股淨值的影響，實證結果分為三個部份；第一部份利用因素分析法，將23個壽險業財務業務變數彙整為4個能力指標，分別是經營能力指標、獲利能力指標、財務結構指標、償債能力指標，第二部份使用集群分析法根據所有壽險公司五年間的績效表現分為5個群集，分析結果顯示不同群集的4個指標能力有顯著差異。分群結果顯示研究期間內各年度的公司表現，其中全球人壽與中國信託人壽在初年度表現非常亮眼，而在後續年度皆保持穩定成長，可能原因在於為了拓展市佔率而導致後續年間的績效表現不如以往亮眼但仍穩定成長。第三部份利用迴歸分析，探討金控公司與非金控公司4個能力指標對公司每股淨值的影響，並建構金控公司與非金控公司的迴歸式，比較兩者之間的差異，結果顯示金控公司與非金控公司適用同一條迴歸式，不論在經營能力或是獲利能力及財務結構表現上，對於公司每股淨值的影響力均相同。[[abstract]]Inthisresearch,24lifeinsurancecompaniesinTaiwanfrom2011to2015areanalyzedforexploringtheeffectofoperationalperformanceonnetassetvaluepershare.Empiricalresultsconsistof3parts.First,factoranalysisisconductedtoclassify23financialvariablesoflifeinsurancecompaniesinto4abilityindicators,Operation,Profitability,FinancialStructureandDebt-Paying.Second,clusteranalysisisimplementedtoclassify5-yearperformanceofalllifeinsurancecompaniesinto5groups.Theresultshowsthat4abilityindicatorsaresignificantlydifferentinthesegroups.Italsoshowsallyearlyperformanceswithinthedefinedperiodofthestudy.TransGlobeLifeandCTBCLifeInsurancebothhaveoutstandingperformanceintheearlyperiodofthestudyandgrowstablyinthelateryears.Thepursuitofmarketsharemaycausethemdecliningperformance,buttheystillhavestablegrowth.Third,usingregressionanalysisexplorestheeffectof4abilityindicatorsforfinancialandnon-financialholdingcompaniesonnetassetvaluepershareandbuildsaregressionmodeloffinancialandnon-financialholdingcompaniesforcomparison.Theresultrevealsthatfinancialandnon-financialholdingcompaniescanusethesameregressionmodelandabilityindicatorsincludingOperation,Profitability,andFinancialStructureproducethesameeffectonnetassetvaluepershare
The Impact of Unemployment Rate and Gross Domestic Product on Life Insurance Premium Income - A Case Study of Asia - Pacific
No full text[[abstract]]本文是使用追蹤資料來分析壽險保費收入與總體經濟變數之間的關係，研究對象為亞太地區，國家數量為19國，時間區間為西元2007-2015年。在實證一開始，我們使用追蹤資料單根檢定，使用的檢定為LLC檢定，該檢定是由Levinetal(2002)提出。根據單根檢定結果，本研究之變數型態為I(0)及I(1)混合序列。之後我們使用Hausman檢定選擇隨機或固定模型，實證結果顯示本文中應選擇隨機模型進行分析。在隨機模型下之實證分析顯示出，壽險保費收入與國內生產毛額(GDP)呈現正相關；而壽險保費收入與失業率則是呈現負相關。[[abstract]]Thispaperappliesthepanelregressionsoflifeinsurancepremiumsandunemploymentrateandgrossdomesticproduct(GDP)for19countriesfrom2007through2015.First,weapplytheLevin,LinandChu(LLC)testdevelopedbyLevinetal.(2002).OurempiricalresultsrevealthatthevariablesareamixtureofI(0)andI(1);second,weusetheHausmanspecificationtestdevelopedbyHausman,J.A.(1978),andtheresultsrevealthatweshouldchoosetherandommodel;finally,thepanelregressionsresultsindicatethatifthegrossdomesticproduct(GDP)raisesthenthelifeinsurancepremiumswillincrease,andiftheunemploymentraisesthenthelifeinsurancepremiumswilldecrease
Optical design of curved light guide plate for desk lamp illumination
No full text[[abstract]]導光板常應用於螢幕顯示器的背光模組，因其具有均勻與低眩光的特性，近年來開始有產品利用導光板於室內照明的應用，但大多都為平板設計，故本文探討弧型導光板於室內照明之光學設計與應用，跳脫以往平板導光元件的刻板外型，在設計弧型導光元件時探討曲率對於整體出光的影響，不同曲率得需要相對應的結構設計做變化，再?用弧型導光板上方微結構設計與分佈達到均勻的光照?設計，增添外觀設計美感與照明產品應用的創新性。在探討照明應用上，則需考?結構對於整體配光曲線的影響光學設計。本文所提出的弧型、多弧型導光板設計於導光板的?側短邊設置LED，藉由頂角為80°的等角V-cut特徵以獲得軸向出光的特性。由於弧型導光板的受照處與特徵分佈不易求得，本文以克??代?模型配合遺傳演算法進?特徵大小分佈的最佳化，提升照?均齊性。本文設計出微笑、飛鳥型與波?型導光板光學設計，具有高照度均勻性，可應用於檯燈與情境照明燈設計，最後並製作原型微笑導光元件與檯燈，並與現有產品比較，驗證設計的優越性及可?性。[[abstract]]Lightguideplates(LGPs)areoftencrutialcomponentsinthebacklightmoduleofliquidcrystaldisplay.Becauseofitsuniformityandlowglarecharacteristics,LGPsbecomepopularinthelightingapplicationsinrecentyears.However,mostLGPsadoptaflatpaneldesign.Thisstudyaddressestheopticaldesignofcurvedlightguideplateofilluminativeuniformity,whichgrealyincreasesthepossibilityforaestheticandinnovativeproductdesignfortheilluminationapplicationsusing,variousarc-shapedLGPs.ThedesignandthedistributionofthecouplingfeatureandtheinfluenceoftheLGPcurvaturetotheopticaldesignarethemajorissuesinthisstudy.Thegeometricaldesignofthemicrofeaturewilldeterminethelightdistributioncurve,andthefeaturedistributionwillinfluenceilluminationuniformity.ThecurvedLGPsstudiedinthethesisincludesasinglearcandmultiple-arcdeisgnofLGPs.TheLEDsarearrangedontheshortedgesofLGPstopreventpossibleglareproblems.Thev-cutwiththevertexangleof80?isselectedduetoapeakintensityattheaxialdirection.Thepitchofv-cutisfixedat2(mm)andthesizeofthev-cutisvariedtoobtaintheilluminationuniformity.Becausethereisnodiectrelationshipbetweentheregionalv-cutdistributionandthelocalilluminationintensityontheobservationsurfaceforcurvedLGPs,thisstudyappliesasurrogateoptimizationusingKrigingmodeltoobtainthesizedistributionofv-cutsforilluminationuniformity.TheopticaldesignofthreecurvedLGPsincludingasinglearc,athree-arc,andamultiple-arctypeshavebeenproposed,whichcanbeappliedtothedesklampswithasmiling,adished,andawavedshapedesign.Aprototypeofasmilingdesklampisfabricatedandtheopticalperformanceismeasuredtocomparewithacommercialproducttoverifythesuperiorityandfeasibilityoftheproposeddesign
A Study on Solar Power Efficiency：Big Data Analysis
No full text[[abstract]]紙本論文延後公開至2022年09月04日止[[abstract]]紙本論文延後公開至2022年09月04日