An Error Detecting and Tagging Framework for Reducing Data Entry Errors in Electronic Medical Records (EMR) System
|
|
- Kevin Sullivan
- 8 years ago
- Views:
Transcription
1 201 IEEE Internatonal Conference on Bonforatcs and Boedcne An Error Detectng and Taggng Fraework for Reducng Data Entry Errors n Electronc Medcal Records (EMR Syste Yuan Lng, Yuan An College of Coputng and Inforatcs, Drexel Unversty Phladelpha, USA {yl68, ya45}@drexel.edu Abstract we develop an error detectng and taggng fraework for reducng data entry errors n Electronc Medcal Records (EMR systes. We propose a taxonoy of data errors w ree levels: Incorrect Forat and Mssng error, Out of Range error, and Inconsstent error. We a to address e challengng proble of detectng erroneous nput values at look statstcally noral but are abnoral n edcal sense. Detectng such an error needs to take patent edcal hstory and populaton data nto consderaton. In partcular, we propose a probablstc eod based on e assupton at e nput value for a feld depends on e hstorcal records of s feld, and s affected by oer felds rough dependency relatonshps. We evaluate our eods usng e data collected fro an EMR Syste. The results show at e eod s prosng for autoatc data entry error detecton. Keywords Data Entry Errors; Electronc Medcal Records (EMR; Error Detectng; Error Taggng I. INTRODUCTION Currently, large-scale data analyss s conducted n alost every area. Advanced analytcal process and data nng technques have been used to dscover valuable nforaton and knowledge fro large-scale data. However, ost data nng technques assue at data has already been cleaned [1]. Ths s not always true n real stuatons. Data preparaton process and data qualty ssues are crtcal to relable data nng results, because data w errors wll coprose e credblty of e data nng results. The data lfecycle of an applcaton typcally conssts of e followng steps: data collecton, transforaton, storage, audtng, cleanng, and analyzng to akng decsons [2]. Data errors can be reduced n each step of e data lfecycle. Snce data collecton s e frst step of e lfecycle, reducng data entry errors n e data collecton stage s e earlest opportunty to enable data qualty. Error and outler detecton have been studed n varous applcaton doans, such as credt card fraud detecton[], network ntruson detecton[4], and data cleanng n data warehouse[5]. In general, errors can be classfed nto ree types: pont errors, contextual errors, and collectve errors [6]. In e healcare doan, data errors n electronc edcal databases have been studed recently [7, 8]. Generally, Mengwen Lu, Xaohua Hu College of Coputng and Inforatcs, Drexel Unversty Phladelpha, USA {l94, xh29}@drexel.edu Electronc Medcal Records (EMR data errors have been classfed as ncoplete, naccurate, or nconsstent [9]. More specfcally, based on EMR st notes, e preous study [10] classfes e data entry errors nto Inconsstent Inforaton, Mscellaneous Errors, Mssng Sectons, Incoplete Inforaton, and Incorrect Inforaton [10]. For our case, we conducted a coprehensve lterature reew. As a result, we suarzed dfferent levels of data errors n a ore detaled anner. Three levels of data entry errors are proposed as shown n Table I. Exaple errors are dscussed n e followng paragraphs. TABLE I. Level 1: Incorrect Forat and Mssng Level 2: Out of Range Level : Inconsstent LEVELS OF DATA ENTRY ERRORS L1-1: Incorrect Forat Data L1-2: Mssng Data L2-1: Out of noral range L2-2: Out of populaton trends L2-: Out of personal trend L-1: Personal Inconsstent L-2: Populaton Inconsstent L1-1 errors can be easly controlled w pror knowledge about e data forat. Constrants can be set for each feld. For exaple, only nubers are allowed to be entered nto blood pressure feld. L2-1 error also can be easly controlled as long as ere are knowledge about e noral range of each feld. For exaple, heght values can be constraned n e range of 0 to eters. Detectng L1-2 errors, however, s a dffcult research task, but we can learn to detect ssng data accordng to clncal gudelnes and caregng practces. Once e ssng values are detected, we can set constrants to s partcular feld as we dd for L1-1 and L2-1. For L2-2 errors, usng statstcs for detectng erroneous values for a sngle attrbute s well-studed and robust results can be obtaned. For exaple, L2-2 errors can be calculated usng a unvarate outler detecton technque called Hapel X84 [11]. In s paper, we do not focus on solng e L1-1 to L2-2 error probles. The challenge les n e detecton of values at look statstcally noral but are abnoral n edcal sense lke L2-, L-1, and L-2 errors. For exaple, e weght value of a patent ncreases to 20lbs fro 180lbs n ree ons. The weght value of 20lbs s n e noral populaton dstrbuton /1/$ IEEE 249
2 range, but for s partcular patent, t s suspcous f he/she has a healy lng style and does not have ajor edcal probles. It s categorzed as a L2- error. The followng llustrates oer exaples. A record showng a ale patent has e status of pregnancy would be consdered as a L-1 error. An obese and dabetc patent w BMI (Body ass ndex larger an 5 s lkely to have a cardovascular dsease anfested as hypertenson and dyslpdea. If e patent s blood pressure values are sgnfcantly low, but stll n e noral range, e blood pressure value would be nconsstent w e BMI value fro e populaton perspectve. Ths sgnals a L-2 type of error. In s paper, we propose an error detectng and taggng echans n real te to address ese challenges based on hstorcal data and dependency relatonshps. The entre fraework conssts of ree an coponents: 1 Fnd a slar patent group. Error detecton s based on hstorcal datasets. For a sngle patent, e hstorcal datasets should be coposed fro patents who have equvalent edcal stuatons, clncal condtons, and basc deographc nforaton. Because t s dffcult to fnd a group of patents w exactly e sae edcal condtons, we detect errors based on a dataset fro patents w slar condtons. 2 Learn a probablstc odel. Based on e dataset fro a group of slar patents, we learn paraeters for our proposed odel. The probablstc odel s used to caculate e nput value probablty for each feld. Return a tag. Error taggng s based on results fro e probablstc odel. If an nput value probablty for a feld s relatvely hgh, en e value has a hgher chance to be a correct value and a Noral tag wll be returned. Whereas f e probablty s relatvely low, e tag s Suspcous. We use a reshold to decde whch tag should be returned. Our experental results show at our error detectng and taggng eod s prosng. The predctve accuracy of e probablstc odel s between 70% and 85% usng our experental data (explaned n e Evaluaton secton. We also copare our eod w exstng eods. The results ndcate at our eod has a better perforance. In suary, e contrbutons of our paper are: 1 We suarze dfferent levels of data entry errors based on a coprehensve lterature reew. Our eod focuses on detectng and taggng errors at cannot be detected by sple statstcal eods. 2 We desgn a probablstc odel based on two assuptons: frst, e current patent s nput value for a feld depends on hs/her hstorcal records n s feld; and second, e nput value for a feld depends on values fro oer felds.we cobne e nfluences fro ese two aspects n our odel. We develop an error detectng and taggng fraework w fve stages: selectng a classfcaton odel, fndng slar patent group, buldng a probablstc odel, settng a reshold, and returnng tags. 4 We test our eod on dataset fro an EMR syste. Our experental results show at e predctve accuracy of our probablstc odel s better an Bayesan Network. Our odel also generates better results an e oer eod assung e nput value for a feld s based solely on hstorcal records. The rest of e paper s organzed as follows: Secton II descrbes related work. Secton III presents our error detectng and taggng fraework. Secton IV ntroduces our probablstc odel and error detectng eod. Secton V presents our evaluaton process and experental results. Secton VI presents our conclusons and future research drectons. II. RELATED WORK Statstcal eods have been used to analyze and detect data errors at data cleanng stage [12-14]. However, we need soe eods to reduce date entry error n e data collecton stage[15]. There are dfferent ways to detect errors and control data qualty, ncludng classfcaton based neural networks, Bayesan networks[16], support vector achne, rule-based eods, unsupersed clusterng, and nstance-based eods. The Usher syste s developed to detect errors on for entry felds by usng Bayesan network and a graphcal odel w explct error odelng[17]. The syste ncludes a probablstc error odel based on a Bayesan network for estatng contextualzed error lkelhood for each feld on a for. The eod uses Bayesan network to learn dependency relatonshps aong felds, but t does not take hstorcal records for a partcular feld nto consderaton. In s paper, we detect errors based on dependency relatonshps aong felds as well as hstorcal data for a partcular feld. In addton, we classfy patents based on soe basc nforaton to cluster e nto slar groups before we apply error detecton. As our eod s probablstc n nature, selectng e rght dataset for estatng e probabltes s crtcal. For exaple, t s unreasonable to predct e nput value for a person whose age s 50 usng a dataset for people whose ages are around 10. Usng taggng echans has been shown to decrease certan types of errors [18]. Data entry nterfaces can be desgned to help prong data qualty. We propose a syste at classfes e nput values of each feld as Suspcous or Noral, and prodes s nforaton as feedback to e user. If e Suspcous tag occurs, en t rends e user to check e nput value. Consequently, t would help reducng data entry errors. III. ERROR DETECTING AND TAGGING FRAMEWORK A. Error Detectng and Taggng Fraework Our error detectng and taggng fraework s presented n Fg. 1. A data entry nterface contans ultple felds. After a user nputs a value for a feld, a tag ( Noral or Suspcous wll be returned. 250
3 probablstc odel and e reshold settng, we return a Noral or Suspcous tag to e current user. IV. DATA ENTRY ERROR DETECTING METHOD Fg. 1. Error Detectng and Taggng Fraework Fg. 2. Hstrocal Records for Error Detectng Task A. Data Entry Error Detectng Meod Gven a data entry nterface I, let F { Fv,, Fj,, Fk } be a set of nput felds on e nterface I. For exaple, n Fg. 4, e feld for enterng e Weght value can be represented as F Weght. Here we use a sple exaple to llustrate e hstorcal data forat and e error detectng task. Fg. 2 shows a set of records for dfferent patents. There are fve felds for each record: Date (Observaton Date, BP_S (Systolc Blood Pressure value, BP_D (Dastolc Blood Pressure value, BMI, and Weght. Our error detectng task s to tell wheer e patent w ID has a noral or suspcous nput value for BP_S, BP_D, BMI and Weght at e observaton date: 1/1/2010. The feedback s generated based on e forer 8 records fro e patent w ID, and also e oer 11 records fro patents w ID 1 and ID 2, assung ese ree patents have slar condtons. Based on e data, we wll learn dependency relatonshps aong BP_S, BP_D, BMI and Weght. These relatonshps are captured n a probablstc graphcal odel. B. Error Detectng Process Based on our error detectng and taggng fraework, we propose e error detectng process as shown n Fg.. A patent s odeled as a set of data felds descrbng e patent s deographc nforaton and clncal condtons. A for s used to collect a set of specfc types of nforaton durng an encounter, for exaple, blood pressure, weght, etc. For our eod, we dde e entre patent nforaton nto two categores: ndependent felds whch are consdered as fxed durng e error detecton process and dependent feld whch wll be collected and flagged as noral or suspcous. Fg.. Error Detectng Process As shown n Fg., our error detectng process conssts of e followng fve steps. Frst, patent records are categorzed nto groups based on nforaton about e basc ndependent felds fro e current patent and oer hstorcal records usng a classfcaton odel. Second, we fnd a slar patent group for e current patent. Thrd, we buld a probablstc odel based on clncal felds and hstorcal records fro a group of slar patents. Four, we set a reshold for e probablty of nput value. Fnally, based on results fro e Fg. 4. A Data Entry For Fg. 5. An exaple of relatonshps aong felds For a sngle feld F v, we can use a sequence of values v { v1,, v 1, } to represent all hstorcal records for a patent, where v s e current value entered by a user for e feld F v, { v 1,, v 1} s e hstorcal record sequence for e feld F and v v s e value entered for e feld 1 F v at e ( 1 te. We assue at a patent s current nput value for a feld depends on hstorcal nput records of s feld. Abnoral changes of nput values coparng w hstorcal records need to get attenton. In probablty eory, such a sequentally dependent process can be descrbed by a Markov odel. Accordng to 1 order Markov assupton, e current value for e feld F v at te depends only on e value for e sae feld at( 1 te. Therefore, for a gng nput value v, we can use e value v to predct e probablty of e value 1 for e feld F v at e current te. The probablty can be represented as q ( 1. Slarly, we can have a probablty accordng to 2 order Markov assupton as q ( 1, 2. But for a new patent, ere are no hstorcal records for value of feld F v. So we can assue at e current value v for e feld F v at te does not depend on preous hstorcal data fro e patent. Therefore e predcted probablty can be represented as q v. ( Our goal s to predct e probablty of an nput value v for e feld F v, whch can be represented as q M ( v. Usng 1 order Markov assupton, 2 order Markov assupton, and a pror probablty, we can calculate q v as n Equaton (1: M ( q M ( v all _ condtons α q ( v 1, v 2, all _ condtons + α 2 q ( v v 1, all _ condtons + α q ( v all _ condtons 1 (1 251
4 all_condtons eans at e probablty s calculated based on a dataset only fro patents at have equal deographc and clncal condtons. In e followng equatons, we wll ot all_condtons n each ter assung at ey have been taken nto consderaton. The ree paraeters α, 1 α, and 2 α are used to balance e ree probabltes [24], andα 1 + α2 + α 1. Gven a set of records, we can estate each part of probablty accordng to Equaton (2, Equaton (, and Equaton (4, respectvely. q ( v v q ( v v q ( v, v all 2, 1, v, v 1, v For Equaton (2 and Equaton (, N ( v 2, 1 s e occurrence nuber of nput value v for feld 2 F at v e ( 2 te followed by e nput value v for feld 1 F v at e ( 1 te for a patent. Slarly, 2, 1, s e occurrence nuber of nput value n a sequence of { 2, 1, } for e ( 2, ( 1, and te for a patent. For Equaton (4, all s e total nuber of values for feld F v for hstorcal records fro all slar patents. Slar patents eans a group of patents who are slar to e current patent. The slarty s easured n ters of er deographc and clncal condtons such as age, gender, dsease condtons, edcatons, treatents, socal acttes, self-anageent, etc. N ( v s e occurrence nuber of v for feld F v fro all hstorcal records based on all slar patents. The Equaton (1 only uses e hstorcal records fro a sngle feld. However, ere are dependency relatonshps aong dfferent felds on a data entry nterface I. The relatonshps aong dfferent felds can be obtaned based on pror knowledge fro experts or learned fro hstorcal data [17]. The relatonshps can be learned fro hstorcal records w a Bayesan network usng e Max-Mn Hll-Clbng (MMHC eod [19]. We pleented e eod usng e R packages: bnlearn [20] and gran [21]. An exaple of learned relatonshps aong felds s showed n Fg. 5. A Condtonal Probablty Table (CPT wll also be learned for e Bayesan Network. For e exaple shown n Fg. 5, e value v of F depends v on e value F w and F u, so we can represent e probablty as q B( w, u. By addng e probablty of q B( w, u to e Equaton (1, we get e followng Equaton (5: q v β q ( v + γ q ( v w, u (5 ( M B where e two paraeters β andγ, β + γ 1. We already get e value of q M ( v by Equaton (1. We can get e value of q v w, u fro CPT of e learned Bayesan B ( (2 ( (4 Network as Fg. 5. By pluggng n e above values, we copute a fnal probablty q( v all _ condtons. B. Error Taggng Meod We assue at for each feld F v e set of possble values s fnte and dscrete. If e values for a feld are contnuous, we apply approaprate eods for dscretzng e. Let ' '' V { v, v, } be e set of possble values for a feld F v. The current nput value v at e te s a eber of e set. For each possble nput values n e set V, we can calculate probablty based on e Equaton (5. We get a tag result for current nput value v for e feld F v as Equaton (6. Noral tag( Suspcous q( θ q( v < θ For e reshold value θ n e Equaton (6, ere are eods at can be appled to set e reshold. For exaple, 80/20 rule, Lnear Utlty Functon [22] and Maxu Lkelhood Estaton[2]. In s paper, we wll use 80/20 rule to dynacally decde e reshold value. In partcular, probablty locates before 20% hghest values eans e correspondng value s Noral, oerwse t s Suspcous. V. EVALUATION The preous sectons descrbed e eoretcal dervatons of e odel and coputatonal steps. In s secton, we present an experent usng a set of lted data for evaluatng e perforance of e odel. We aed to gan soe frsand practcal confdence rough a plot study. The ltaton of e data s two-fold. Frst, we only obtaned e data for a sall nuber of people, about Second, we consder a patent odel w a sall nuber of felds ncludng age, sex, BP, weght, and BMI. The experent s by no eans orough and coplete because e real odel for descrbng a patent s far ore coplcated an just sevaral felds. The purpose of e experent s to deonstrate e effectveness of e probablstc odel for error detecton takng hstorcal data and nter-feld dependency relatonshps nto consderaton. We obtaned e data fro a heal clnc afflated w Drexel Unversty n Phladepha, PA, USA. All e data are de-dentfed for HIPAA (Heal Insurance Portablty and Accountablty Act coplance. We only obtaned records w sx felds wout any addtonal clncal nforaton. A. Data Sets We selected sx felds fro databases: Age, Sex, Dastolc Blood Pressure (BP_D value, Systolc Blood Pressure (BP_S value, Weght, and Body Mass Index (BMI. we extracted 1448 patents w records fro all ese sx felds fro e EMR syste. Because our dataset s lted and patents nforaton s nsuffcent, we cannot guarantee all patents have equal clncal condtons and treatents n s sall scale prelnary study. Only Age and Sex are used as basc felds to conduct classfcaton process to fnd a slar patents group for current patent. Based on e two basc attrbutes: age and sex. Patents are dded nto 14 categores and e nubers of (6 252
5 patents for each category are shown n Table II. The average nuber of hstorcal records for each patent s 8.4. We exane our data entry error detectng eod for each group. TABLE II. 16 KINDS OF SIMILAR PATIENTS Patents/Records Group Age Sex Nuber 1 M 0/15 [20, 0 2 F 6/567 M 68/28 [0, 40 4 F 12/97 5 M 122/64 [40, 50 6 F 188/ M 191/151 [50, 60 8 F 270/ M 10/850 [60, F 181/ M 20/195 [70, F 55/658 1 M 4/6 [80, F 21/172 a. M stands for Male, and F stands for Feale BP_D, BP_S, Weght, and BMI are transfored to dscrete value and used as clncal felds to buld a Bayesan Network to fnd e nter-feld dependency relatonshp aong ese clncal felds. In our experent, we bult a unfed Bayesan network usng all records. The Max-Mn Hll-Clbng (MMHC eod s used to generate e Bayesan Network structure as shown n Fg. 6. Fg. 6. Learned Bayesan Network Fg. 7. Predctve Accuracy and Nuber of Records of 14 Groups The varable BP_D depends on e varable BP_S. The jont probablty between BP_D and BP_S can be calculated fro e CPT (Condtonal Probablty Table of Bayesan Network. B. Experental Results Based on our dataset, we used our eod to predct e value of varable BP_D. The predctve power of our eod s easured n nddual -fold cross valdaton experents for each group. The average predctve accuracy and nuber of records for each group s shown n Fg. 7. The accuracy s calculated as correctly predcted nstance nuber dded by total nstance nuber. The paraeters set here are α α α 1/, β / 4, and γ 1/ Fg. 8. Coparatve Perforance of Usher w Our Meod Fg. 9. Accuracy Coparson for Dfferent Paraeters Set The average predctve accuracy for all groups s 71%. The predctve power s stable for all ese groups, whch ranges fro 70% to 85%, except for group1, group1, and group14. Ther nubers of records are lower an 200, so er predctve accuracy are relatvely lower. We pleent Usher s error odel [17] on our dataset. Usher nduced e probablty of akng an error usng an error odel. The coparatve results of Usher and our eod s shown n Fg. 8. Usher s average accuracy s 6.2%, whle our eod s 71%. Except for group 1 and group 1 n whch e nubers of records are too sall, our eod has better perforance an Usher. Snce e nuber of records and accuracy of group 8 s relatve hgh. We test dfferent sets of paraeters on e dataset of group 8. The dfferent paraeter sets and correspondng accuracy results are as shown as Table III and Fg. 9. There are 9 dfferent sets of paraeters. TABLE III. ACCURACY RESULTS FOR DIFFERENT PARAMTERS SET Set α1 α α 2 β γ Accuracy / 1/ 1/ / 1/ 1/ / 1/ 1/ As shown n Table III, n Set 1 and Set 2, β 1 andγ 0, whch eans at predcton only depends on past hstorcal values of e current sngle feld. In set, β 0 andγ 1, whch eans at predcton only depends on oer felds, but does not depend on hstorcal values of current sngle feld. That s, e result of set s e result of only usng Bayesan network. The accuracy results fro set 1, 2 and are relatvely low. In suary, e accuracy results obtaned by e eods purely dependng on past hstorcal values of e current sngle feld or on dependency relatonshps aong ultple felds learned fro Bayesan network are not as good as e results obtaned by cobnng ese two aspects. In Fg. 9, sets 6, 7, 8, and 9 have relatvely hgher accuracy. In ese four sets, paraeters β and α have hgher weght. That eans predcton dependng on all forer 2 stages of hstorcal values of e current sngle feld and dependency relatonshps aong ultple felds proves e accuracy. The set 9 has hghest accuracy. But n practce, t s not recoended, because for soe sall datasets, α ght equal to 0. The recoended paraeter sets are Set 6, 7, and 8. We also conduct an experent to copare e perforance of our odel to at of e Hapel X84[11]. Based on our eod, e Noral ranges for a patent s BP_D value can dynacally narrow down to [70, 90] as shown n Fg. 10. a, whle e Hapel X84 gave Noral to a broader range [45., 114.7] as shown n Fg. 10. b for all populaton. The narrower range ndcates at our eod s ore senstve to error detecton. 25
6 a Our eod b Hapel X84 Fg. 10. Our eod copared w Hapel X84 VI. DISCUSSION AND FUTURE WORK Ths paper proposes a fve-step error detectng and taggng fraework for reducng data entry errors at pose sgnfcant challenges to exstng statstc eods. Our probablstc odel for error detectng s based on e assupton at nput values for a feld can be predcated usng hstorcal data and nter-feld dependency relatonshps aong felds. We conducted plot experents to evaluate our eod. Results show at our probablstc odel has better accuracy an a Bayesan network, t s also better an e eod based on a Markov assupton at e nput value for a feld s based only on hstorcal records of s sngle feld. We also copare our eod w Usher s error odel. The result shows our eod has a better perforance. Moreover, our eod has ore precse results an e Hapel X84 eod. Our future research work wll focus on followng aspects: 1 Currently, our taggng eod can only return two knds of tags as Noral and Suspcous. We would lke to extend our odel to detect ultple errors as shown n Table I. 2 Our error detectng eod only focuses on dscrete nput value. We plan to prove our eod to deal w contnuous nput values. We plan to conduct a ore orough experental evaluaton by collectng ore patent data w ore attrbutes n e future. ACKNOWLEDGMENT Ths work s supported n part by a Drexel Jupstart grant on Heal Inforatcs and e NSF grant IIP for e center for sual and decson nforatcs (CVDI. REFERENCES [1] Dasu, T. and T. Johnson, Exploratory data nng and data cleanng. Vol : Wley-Interscence. [2] Batn, C. and M. Scannapeca, Data qualty. 2006: Sprnger. [] Chan, P.K., et al., Dstrbuted data nng n credt card fraud detecton. Intellgent Systes and er Applcatons, IEEE, (6: p [4] Lazarec, A., et al., A coparatve study of anoaly detecton schees n network ntruson detecton. Proc. SIAM, 200. [5] Rah, E. and H.H. Do, Data cleanng: Probles and current approaches. IEEE Data Eng. Bull., (4: p. -1. [6] Chandola, V., A. Banerjee, and V. Kuar, Anoaly Detecton: A Survey. Ac Coputng Surveys, (. [7] Arts, D.G., N.F. De Kezer, and G.-J. Scheffer, Defnng and prong data qualty n edcal regstres: a lterature reew, case study, and generc fraework. Journal of e Aercan Medcal Inforatcs Assocaton, (6: p [8] Beretta, L., et al., Iprong e qualty of data entry n a low-budget head njury database. Acta Neurochrurgca, (9: p [9] Phllps, W. and Y. Gong, Developng a noenclature for er errors, n Huan-Coputer Interacton. Interactng n Varous Applcaton Doans. 2009, Sprnger. p [10] Khare, R., An, Y., Wolf, S., Nyrjesy, P., Lu, L., & Chou, E, Understandng e EMR error control practces aong gynecologc physcans. Conference 201 Proceedngs, p [11] Hellersten, J.M., Quanttatve data cleanng for large databases. Unted Natons Econoc Cosson for Europe (UNECE, [12] Nah, M.L., C.F. Peper, and M.M. Cunnngha, Quantfyng Data Qualty for Clncal Trals Usng Electronc Data Capture. Plos One, (8. [1] Goldberg, S.I., A. Neerko, and A. Turchn, Analyss of data errors n clncal research databases. AMIA Annu Syp Proc, 2008: p [14] Jacobs, B., Electronc edcal record, error detecton, and error reducton: a pedatrc crtcal care perspectve. Pedatr Crt Care Med, (2 Suppl: p. S [15] Goldberg, S.I., et al., A Weghty Proble: Identfcaton, Characterstcs and Rsk Factors for Errors n EMR Data. AMIA Annu Syp Proc, : p [16] Wong, W., et al. Bayesan network anoaly pattern detecton for dsease outbreaks. n MACHINE LEARNING-INTERNATIONAL WORKSHOP THEN CONFERENCE [17] Chen, K., et al., Usher: Iprong data qualty w dynac fors. Knowledge and Data Engneerng, IEEE Transactons on, (8: p [18] Chen, K., J.M. Hellersten, and T.S. Parkh. Desgnng adaptve feedback for prong data entry accuracy. n Proceedngs of e 2nd annual ACM syposu on User nterface software and technology ACM. [19] Tsaardnos, I., L.E. Brown, and C.F. Alfers, The ax-n hllclbng Bayesan network structure learnng algor. Machne learnng, (1: p [20] Scutar, M., Learnng Bayesan networks w e bnlearn R package. arxv preprnt arxv: , [21] Højsgaard, S., Graphcal Independence Networks w e gran package for R, 2012, Journal. [22] Arapatzs, A., et al., KUN on e TREC-9 Flterng Track: Increentalty, decay, and reshold optzaton for adaptve flterng systes [2] Zhang, Y. and J. Callan. Maxu lkelhood estaton for flterng resholds. n Proceedngs of e 24 annual nternatonal ACM SIGIR conference on Research and developent n nforaton retreval ACM. [24] Mchael Collns. Language Modellng, llns/
An Electricity Trade Model for Microgrid Communities in Smart Grid
An Electrcty Trade Model for Mcrogrd Countes n Sart Grd Tansong Cu, Yanzh Wang, Shahn Nazaran and Massoud Pedra Unversty of Southern Calforna Departent of Electrcal Engneerng Los Angeles, CA, USA {tcu,
More informationForecasting the Direction and Strength of Stock Market Movement
Forecastng the Drecton and Strength of Stock Market Movement Jngwe Chen Mng Chen Nan Ye cjngwe@stanford.edu mchen5@stanford.edu nanye@stanford.edu Abstract - Stock market s one of the most complcated systems
More informationA Statistical Model for Detecting Abnormality in Static-Priority Scheduling Networks with Differentiated Services
A Statstcal odel for Detectng Abnoralty n Statc-Prorty Schedulng Networks wth Dfferentated Servces ng L 1 and We Zhao 1 School of Inforaton Scence & Technology, East Chna Noral Unversty, Shangha 0006,
More informationbenefit is 2, paid if the policyholder dies within the year, and probability of death within the year is ).
REVIEW OF RISK MANAGEMENT CONCEPTS LOSS DISTRIBUTIONS AND INSURANCE Loss and nsurance: When someone s subject to the rsk of ncurrng a fnancal loss, the loss s generally modeled usng a random varable or
More informationInstitute of Informatics, Faculty of Business and Management, Brno University of Technology,Czech Republic
Lagrange Multplers as Quanttatve Indcators n Economcs Ivan Mezník Insttute of Informatcs, Faculty of Busness and Management, Brno Unversty of TechnologCzech Republc Abstract The quanttatve role of Lagrange
More informationForecasting the Demand of Emergency Supplies: Based on the CBR Theory and BP Neural Network
700 Proceedngs of the 8th Internatonal Conference on Innovaton & Management Forecastng the Demand of Emergency Supples: Based on the CBR Theory and BP Neural Network Fu Deqang, Lu Yun, L Changbng School
More informationTwo-Phase Traceback of DDoS Attacks with Overlay Network
4th Internatonal Conference on Sensors, Measureent and Intellgent Materals (ICSMIM 205) Two-Phase Traceback of DDoS Attacks wth Overlay Network Zahong Zhou, a, Jang Wang2, b and X Chen3, c -2 School of
More informationHow Much to Bet on Video Poker
How Much to Bet on Vdeo Poker Trstan Barnett A queston that arses whenever a gae s favorable to the player s how uch to wager on each event? Whle conservatve play (or nu bet nzes large fluctuatons, t lacks
More informationA Fuzzy Optimization Framework for COTS Products Selection of Modular Software Systems
Internatonal Journal of Fuy Systes, Vol. 5, No., June 0 9 A Fuy Optaton Fraework for COTS Products Selecton of Modular Software Systes Pankaj Gupta, Hoang Pha, Mukesh Kuar Mehlawat, and Shlp Vera Abstract
More informationBasic Queueing Theory M/M/* Queues. Introduction
Basc Queueng Theory M/M/* Queues These sldes are created by Dr. Yh Huang of George Mason Unversty. Students regstered n Dr. Huang's courses at GMU can ake a sngle achne-readable copy and prnt a sngle copy
More informationCS 2750 Machine Learning. Lecture 3. Density estimation. CS 2750 Machine Learning. Announcements
Lecture 3 Densty estmaton Mlos Hauskrecht mlos@cs.ptt.edu 5329 Sennott Square Next lecture: Matlab tutoral Announcements Rules for attendng the class: Regstered for credt Regstered for audt (only f there
More informationBANDWIDTH ALLOCATION AND PRICING PROBLEM FOR A DUOPOLY MARKET
Yugoslav Journal of Operatons Research (0), Nuber, 65-78 DOI: 0.98/YJOR0065Y BANDWIDTH ALLOCATION AND PRICING PROBLEM FOR A DUOPOLY MARKET Peng-Sheng YOU Graduate Insttute of Marketng and Logstcs/Transportaton,
More informationBayesian Network Based Causal Relationship Identification and Funding Success Prediction in P2P Lending
Proceedngs of 2012 4th Internatonal Conference on Machne Learnng and Computng IPCSIT vol. 25 (2012) (2012) IACSIT Press, Sngapore Bayesan Network Based Causal Relatonshp Identfcaton and Fundng Success
More informationThe Development of Web Log Mining Based on Improve-K-Means Clustering Analysis
The Development of Web Log Mnng Based on Improve-K-Means Clusterng Analyss TngZhong Wang * College of Informaton Technology, Luoyang Normal Unversty, Luoyang, 471022, Chna wangtngzhong2@sna.cn Abstract.
More informationCONSTRUCTION OF A COLLABORATIVE VALUE CHAIN IN CLOUD COMPUTING ENVIRONMENT
CONSTRUCTION OF A COLLAORATIVE VALUE CHAIN IN CLOUD COMPUTING ENVIRONMENT Png Wang, School of Econoy and Manageent, Jangsu Unversty of Scence and Technology, Zhenjang Jangsu Chna, sdwangp1975@163.co Zhyng
More informationAn Alternative Way to Measure Private Equity Performance
An Alternatve Way to Measure Prvate Equty Performance Peter Todd Parlux Investment Technology LLC Summary Internal Rate of Return (IRR) s probably the most common way to measure the performance of prvate
More informationRecurrence. 1 Definitions and main statements
Recurrence 1 Defntons and man statements Let X n, n = 0, 1, 2,... be a MC wth the state space S = (1, 2,...), transton probabltes p j = P {X n+1 = j X n = }, and the transton matrx P = (p j ),j S def.
More informationSIMPLE LINEAR CORRELATION
SIMPLE LINEAR CORRELATION Smple lnear correlaton s a measure of the degree to whch two varables vary together, or a measure of the ntensty of the assocaton between two varables. Correlaton often s abused.
More informationMaximizing profit using recommender systems
Maxzng proft usng recoender systes Aparna Das Brown Unversty rovdence, RI aparna@cs.brown.edu Clare Matheu Brown Unversty rovdence, RI clare@cs.brown.edu Danel Rcketts Brown Unversty rovdence, RI danel.bore.rcketts@gal.co
More informationRisk-based Fatigue Estimate of Deep Water Risers -- Course Project for EM388F: Fracture Mechanics, Spring 2008
Rsk-based Fatgue Estmate of Deep Water Rsers -- Course Project for EM388F: Fracture Mechancs, Sprng 2008 Chen Sh Department of Cvl, Archtectural, and Envronmental Engneerng The Unversty of Texas at Austn
More informationFeature selection for intrusion detection. Slobodan Petrović NISlab, Gjøvik University College
Feature selecton for ntruson detecton Slobodan Petrovć NISlab, Gjøvk Unversty College Contents The feature selecton problem Intruson detecton Traffc features relevant for IDS The CFS measure The mrmr measure
More informationQuality of Service Analysis and Control for Wireless Sensor Networks
Qualty of ervce Analyss and Control for Wreless ensor Networs Jaes Kay and Jeff Frol Unversty of Veront ay@uv.edu, frol@eba.uv.edu Abstract hs paper nvestgates wreless sensor networ spatal resoluton as
More informationA NOTE ON THE PREDICTION AND TESTING OF SYSTEM RELIABILITY UNDER SHOCK MODELS C. Bouza, Departamento de Matemática Aplicada, Universidad de La Habana
REVISTA INVESTIGACION OPERACIONAL Vol., No. 3, 000 A NOTE ON THE PREDICTION AND TESTING OF SYSTEM RELIABILITY UNDER SHOCK MODELS C. Bouza, Departaento de Mateátca Aplcada, Unversdad de La Habana ABSTRACT
More informationHow To Understand The Results Of The German Meris Cloud And Water Vapour Product
Ttel: Project: Doc. No.: MERIS level 3 cloud and water vapour products MAPP MAPP-ATBD-ClWVL3 Issue: 1 Revson: 0 Date: 9.12.1998 Functon Name Organsaton Sgnature Date Author: Bennartz FUB Preusker FUB Schüller
More informationA Novel Dynamic Role-Based Access Control Scheme in User Hierarchy
Journal of Coputatonal Inforaton Systes 6:7(200) 2423-2430 Avalable at http://www.jofcs.co A Novel Dynac Role-Based Access Control Schee n User Herarchy Xuxa TIAN, Zhongqn BI, Janpng XU, Dang LIU School
More informationThe OC Curve of Attribute Acceptance Plans
The OC Curve of Attrbute Acceptance Plans The Operatng Characterstc (OC) curve descrbes the probablty of acceptng a lot as a functon of the lot s qualty. Fgure 1 shows a typcal OC Curve. 10 8 6 4 1 3 4
More informationA Hierarchical Anomaly Network Intrusion Detection System using Neural Network Classification
IDC IDC A Herarchcal Anomaly Network Intruson Detecton System usng Neural Network Classfcaton ZHENG ZHANG, JUN LI, C. N. MANIKOPOULOS, JAY JORGENSON and JOSE UCLES ECE Department, New Jersey Inst. of Tech.,
More informationTHE APPLICATION OF DATA MINING TECHNIQUES AND MULTIPLE CLASSIFIERS TO MARKETING DECISION
Internatonal Journal of Electronc Busness Management, Vol. 3, No. 4, pp. 30-30 (2005) 30 THE APPLICATION OF DATA MINING TECHNIQUES AND MULTIPLE CLASSIFIERS TO MARKETING DECISION Yu-Mn Chang *, Yu-Cheh
More informationAn Analytical Model of Web Server Load Distribution by Applying a Minimum Entropy Strategy
Internatonal Journal of Coputer and Councaton Engneerng, Vol. 2, No. 4, July 203 An Analytcal odel of Web Server Load Dstrbuton by Applyng a nu Entropy Strategy Teeranan Nandhakwang, Settapong alsuwan,
More informationNaglaa Raga Said Assistant Professor of Operations. Egypt.
Volue, Issue, Deceer ISSN: 77 8X Internatonal Journal of Adanced Research n Coputer Scence and Software Engneerng Research Paper Aalale onlne at: www.jarcsse.co Optal Control Theory Approach to Sole Constraned
More informationCalculation of Sampling Weights
Perre Foy Statstcs Canada 4 Calculaton of Samplng Weghts 4.1 OVERVIEW The basc sample desgn used n TIMSS Populatons 1 and 2 was a two-stage stratfed cluster desgn. 1 The frst stage conssted of a sample
More informationA Novel Methodology of Working Capital Management for Large. Public Constructions by Using Fuzzy S-curve Regression
Novel Methodology of Workng Captal Management for Large Publc Constructons by Usng Fuzzy S-curve Regresson Cheng-Wu Chen, Morrs H. L. Wang and Tng-Ya Hseh Department of Cvl Engneerng, Natonal Central Unversty,
More informationAn Enhanced K-Anonymity Model against Homogeneity Attack
JOURNAL OF SOFTWARE, VOL. 6, NO. 10, OCTOBER 011 1945 An Enhanced K-Anont Model aganst Hoogenet Attack Qan Wang College of Coputer Scence of Chongqng Unverst, Chongqng, Chna Eal: wangqan@cqu.edu.cn Zhwe
More informationStochastic Models of Load Balancing and Scheduling in Cloud Computing Clusters
Stochastc Models of Load Balancng and Schedulng n Cloud Coputng Clusters Sva Theja Magulur and R. Srkant Departent of ECE and CSL Unversty of Illnos at Urbana-Chapagn sva.theja@gal.co; rsrkant@llnos.edu
More informationVirtual machine resource allocation algorithm in cloud environment
COMPUTE MOELLIN & NEW TECHNOLOIES 2014 1(11) 279-24 Le Zheng Vrtual achne resource allocaton algorth n cloud envronent 1, 2 Le Zheng 1 School of Inforaton Engneerng, Shandong Youth Unversty of Poltcal
More informationThe Use of Analytics for Claim Fraud Detection Roosevelt C. Mosley, Jr., FCAS, MAAA Nick Kucera Pinnacle Actuarial Resources Inc.
Paper 1837-2014 The Use of Analytcs for Clam Fraud Detecton Roosevelt C. Mosley, Jr., FCAS, MAAA Nck Kucera Pnnacle Actuaral Resources Inc., Bloomngton, IL ABSTRACT As t has been wdely reported n the nsurance
More informationInternational Journal of Industrial Engineering Computations
Internatonal Journal of Industral ngneerng Coputatons 3 (2012) 393 402 Contents lsts avalable at GrowngScence Internatonal Journal of Industral ngneerng Coputatons hoepage: www.growngscence.co/jec Suppler
More informationRobust Design of Public Storage Warehouses. Yeming (Yale) Gong EMLYON Business School
Robust Desgn of Publc Storage Warehouses Yemng (Yale) Gong EMLYON Busness School Rene de Koster Rotterdam school of management, Erasmus Unversty Abstract We apply robust optmzaton and revenue management
More informationIMPACT ANALYSIS OF A CELLULAR PHONE
4 th ASA & μeta Internatonal Conference IMPACT AALYSIS OF A CELLULAR PHOE We Lu, 2 Hongy L Bejng FEAonlne Engneerng Co.,Ltd. Bejng, Chna ABSTRACT Drop test smulaton plays an mportant role n nvestgatng
More informationA Hybrid Approach to Evaluate the Performance of Engineering Schools
A Hybrd Approach to Evaluate the Perforance of Engneerng Schools School of Engneerng Unversty of Brdgeport Brdgeport, CT 06604 ABSTRACT Scence and engneerng (S&E) are two dscplnes that are hghly receptve
More informationScan Detection in High-Speed Networks Based on Optimal Dynamic Bit Sharing
Scan Detecton n Hgh-Speed Networks Based on Optal Dynac Bt Sharng Tao L Shgang Chen Wen Luo Mng Zhang Departent of Coputer & Inforaton Scence & Engneerng, Unversty of Florda Abstract Scan detecton s one
More informationCapacity Planning for Virtualized Servers
Capacty Plannng for Vrtualzed Servers Martn Bchler, Thoas Setzer, Benjan Spetkap Departent of Inforatcs, TU München 85748 Garchng/Munch, Gerany (bchler setzer benjan.spetkap)@n.tu.de Abstract Today's data
More informationPacket Reorderng Analysis
On Montorng of End-to-End Packet Reorderng over the Internet Bn Ye 1 Anura P. Jayasuana 1 Nschal M. Pratla 2 1Coputer Networkng Research laboratory, Colorado State Unversty, Fort Collns, CO 8523, USA 2
More informationAn Interest-Oriented Network Evolution Mechanism for Online Communities
An Interest-Orented Network Evoluton Mechansm for Onlne Communtes Cahong Sun and Xaopng Yang School of Informaton, Renmn Unversty of Chna, Bejng 100872, P.R. Chna {chsun,yang}@ruc.edu.cn Abstract. Onlne
More informationWeb Service-based Business Process Automation Using Matching Algorithms
Web Servce-based Busness Process Autoaton Usng Matchng Algorths Yanggon K and Juhnyoung Lee 2 Coputer and Inforaton Scences, Towson Uversty, Towson, MD 2252, USA, yk@towson.edu 2 IBM T. J. Watson Research
More informationMultiple-Period Attribution: Residuals and Compounding
Multple-Perod Attrbuton: Resduals and Compoundng Our revewer gave these authors full marks for dealng wth an ssue that performance measurers and vendors often regard as propretary nformaton. In 1994, Dens
More informationHow To Calculate The Accountng Perod Of Nequalty
Inequalty and The Accountng Perod Quentn Wodon and Shlomo Ytzha World Ban and Hebrew Unversty September Abstract Income nequalty typcally declnes wth the length of tme taen nto account for measurement.
More informationInternational Journal of Information Management
Internatonal Journal of Inforaton Manageent 32 (2012) 409 418 Contents lsts avalable at ScVerse ScenceDrect Internatonal Journal of Inforaton Manageent j our nal ho e p age: www.elsever.co/locate/jnfogt
More informationA Similar Duplicate Data Detection Method Based on Fuzzy Clustering for Topology Formation
Lejang GUO 1, We WANG 2, Fangxn CHEN 1, Xao TANG 1,2, Weang WANG 1 Ar Force Radar Acadey (1), Wuhan Unversty (2) A Slar Duplcate Data Detecton Method Based on Fuzzy Clusterng for Topology Foraton Abstract.
More informationStochastic Models of Load Balancing and Scheduling in Cloud Computing Clusters
01 Proceedngs IEEE INFOCOM Stochastc Models of Load Balancng and Schedulng n Cloud Coputng Clusters Sva heja Magulur and R. Srkant Departent of ECE and CSL Unversty of Illnos at Urbana-Chapagn sva.theja@gal.co;
More informationWhat is Candidate Sampling
What s Canddate Samplng Say we have a multclass or mult label problem where each tranng example ( x, T ) conssts of a context x a small (mult)set of target classes T out of a large unverse L of possble
More informationA Hybrid Discriminative/Generative Approach for Modeling Human Activities
A Hybrd Dscrnatve/Generatve Approach for Modelng Huan Actvtes Jonathan Lester 1, Tanzee Choudhury 2, Ncky Kern 3, Gaetano Borrello 2,4 and Blake Hannaford 1 1 Departent of Electrcal Engneerng, Unversty
More informationMining Multiple Large Data Sources
The Internatonal Arab Journal of Informaton Technology, Vol. 7, No. 3, July 2 24 Mnng Multple Large Data Sources Anmesh Adhkar, Pralhad Ramachandrarao 2, Bhanu Prasad 3, and Jhml Adhkar 4 Department of
More informationCan Auto Liability Insurance Purchases Signal Risk Attitude?
Internatonal Journal of Busness and Economcs, 2011, Vol. 10, No. 2, 159-164 Can Auto Lablty Insurance Purchases Sgnal Rsk Atttude? Chu-Shu L Department of Internatonal Busness, Asa Unversty, Tawan Sheng-Chang
More informationLinear Circuits Analysis. Superposition, Thevenin /Norton Equivalent circuits
Lnear Crcuts Analyss. Superposton, Theenn /Norton Equalent crcuts So far we hae explored tmendependent (resste) elements that are also lnear. A tmendependent elements s one for whch we can plot an / cure.
More informationBUSINESS PROCESS PERFORMANCE MANAGEMENT USING BAYESIAN BELIEF NETWORK. 0688, dskim@ssu.ac.kr
Proceedngs of the 41st Internatonal Conference on Computers & Industral Engneerng BUSINESS PROCESS PERFORMANCE MANAGEMENT USING BAYESIAN BELIEF NETWORK Yeong-bn Mn 1, Yongwoo Shn 2, Km Jeehong 1, Dongsoo
More informationRevenue Maximization Using Adaptive Resource Provisioning in Cloud Computing Environments
202 ACM/EEE 3th nternatonal Conference on Grd Coputng evenue Maxzaton sng Adaptve esource Provsonng n Cloud Coputng Envronents Guofu Feng School of nforaton Scence, Nanng Audt nversty, Nanng, Chna nufgf@gal.co
More informationDetecting Credit Card Fraud using Periodic Features
Detectng Credt Card Fraud usng Perodc Features Alejandro Correa Bahnsen, Djamla Aouada, Aleksandar Stojanovc and Björn Ottersten Interdscplnary Centre for Securty, Relablty and Trust Unversty of Luxembourg,
More informationCHAPTER 14 MORE ABOUT REGRESSION
CHAPTER 14 MORE ABOUT REGRESSION We learned n Chapter 5 that often a straght lne descrbes the pattern of a relatonshp between two quanttatve varables. For nstance, n Example 5.1 we explored the relatonshp
More informationTraffic Demand Forecasting for EGCS with Grey Theory Based Multi- Model Method
IJCSI Internatonal Journal of Coputer Scence Issues, Vol., Issue, No, January 3 ISSN (Prnt): 694-784 ISSN (Onlne): 694-84 www.ijcsi.org 6 Traffc Deand Forecastng for EGCS wth Grey Theory Based Mult- Model
More informationStochastic Models of Load Balancing and Scheduling in Cloud Computing Clusters
Stochastc Models of Load Balancng and Schedulng n Cloud Coputng Clusters Sva Theja Magulur and R. Srkant Departent of ECE and CSL Unversty of Illnos at Urbana-Chapagn sva.theja@gal.co; rsrkant@llnos.edu
More informationII. THE QUALITY AND REGULATION OF THE DISTRIBUTION COMPANIES I. INTRODUCTION
Fronter Methodology to fx Qualty goals n Electrcal Energy Dstrbuton Copanes R. Rarez 1, A. Sudrà 2, A. Super 3, J.Bergas 4, R.Vllafáfla 5 1-2 -3-4-5 - CITCEA - UPC UPC., Unversdad Poltécnca de Cataluña,
More informationAn Evaluation of the Extended Logistic, Simple Logistic, and Gompertz Models for Forecasting Short Lifecycle Products and Services
An Evaluaton of the Extended Logstc, Smple Logstc, and Gompertz Models for Forecastng Short Lfecycle Products and Servces Charles V. Trappey a,1, Hsn-yng Wu b a Professor (Management Scence), Natonal Chao
More informationOn-Line Fault Detection in Wind Turbine Transmission System using Adaptive Filter and Robust Statistical Features
On-Lne Fault Detecton n Wnd Turbne Transmsson System usng Adaptve Flter and Robust Statstcal Features Ruoyu L Remote Dagnostcs Center SKF USA Inc. 3443 N. Sam Houston Pkwy., Houston TX 77086 Emal: ruoyu.l@skf.com
More informationCredit Limit Optimization (CLO) for Credit Cards
Credt Lmt Optmzaton (CLO) for Credt Cards Vay S. Desa CSCC IX, Ednburgh September 8, 2005 Copyrght 2003, SAS Insttute Inc. All rghts reserved. SAS Propretary Agenda Background Tradtonal approaches to credt
More informationFREQUENCY OF OCCURRENCE OF CERTAIN CHEMICAL CLASSES OF GSR FROM VARIOUS AMMUNITION TYPES
FREQUENCY OF OCCURRENCE OF CERTAIN CHEMICAL CLASSES OF GSR FROM VARIOUS AMMUNITION TYPES Zuzanna BRO EK-MUCHA, Grzegorz ZADORA, 2 Insttute of Forensc Research, Cracow, Poland 2 Faculty of Chemstry, Jagellonan
More informationMethodology to Determine Relationships between Performance Factors in Hadoop Cloud Computing Applications
Methodology to Determne Relatonshps between Performance Factors n Hadoop Cloud Computng Applcatons Lus Eduardo Bautsta Vllalpando 1,2, Alan Aprl 1 and Alan Abran 1 1 Department of Software Engneerng and
More informationMETHODOLOGY TO DETERMINE RELATIONSHIPS BETWEEN PERFORMANCE FACTORS IN HADOOP CLOUD COMPUTING APPLICATIONS
METHODOLOGY TO DETERMINE RELATIONSHIPS BETWEEN PERFORMANCE FACTORS IN HADOOP CLOUD COMPUTING APPLICATIONS Lus Eduardo Bautsta Vllalpando 1,2, Alan Aprl 1 and Alan Abran 1 1 Department of Software Engneerng
More informationStatistical Methods to Develop Rating Models
Statstcal Methods to Develop Ratng Models [Evelyn Hayden and Danel Porath, Österrechsche Natonalbank and Unversty of Appled Scences at Manz] Source: The Basel II Rsk Parameters Estmaton, Valdaton, and
More informationAddendum to: Importing Skill-Biased Technology
Addendum to: Importng Skll-Based Technology Arel Bursten UCLA and NBER Javer Cravno UCLA August 202 Jonathan Vogel Columba and NBER Abstract Ths Addendum derves the results dscussed n secton 3.3 of our
More informationRELIABILITY, RISK AND AVAILABILITY ANLYSIS OF A CONTAINER GANTRY CRANE ABSTRACT
Kolowrock Krzysztof Joanna oszynska MODELLING ENVIRONMENT AND INFRATRUCTURE INFLUENCE ON RELIABILITY AND OPERATION RT&A # () (Vol.) March RELIABILITY RIK AND AVAILABILITY ANLYI OF A CONTAINER GANTRY CRANE
More informationCalculating the high frequency transmission line parameters of power cables
< ' Calculatng the hgh frequency transmsson lne parameters of power cables Authors: Dr. John Dcknson, Laboratory Servces Manager, N 0 RW E B Communcatons Mr. Peter J. Ncholson, Project Assgnment Manager,
More informationTechnical Report, SFB 475: Komplexitätsreduktion in Multivariaten Datenstrukturen, Universität Dortmund, No. 1998,04
econstor www.econstor.eu Der Open-Access-Publkatonsserver der ZBW Lebnz-Inforatonszentru Wrtschaft The Open Access Publcaton Server of the ZBW Lebnz Inforaton Centre for Econocs Becka, Mchael Workng Paper
More informationDEFINING %COMPLETE IN MICROSOFT PROJECT
CelersSystems DEFINING %COMPLETE IN MICROSOFT PROJECT PREPARED BY James E Aksel, PMP, PMI-SP, MVP For Addtonal Informaton about Earned Value Management Systems and reportng, please contact: CelersSystems,
More informationModule 2 LOSSLESS IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur
Module LOSSLESS IMAGE COMPRESSION SYSTEMS Lesson 3 Lossless Compresson: Huffman Codng Instructonal Objectves At the end of ths lesson, the students should be able to:. Defne and measure source entropy..
More informationAN APPOINTMENT ORDER OUTPATIENT SCHEDULING SYSTEM THAT IMPROVES OUTPATIENT EXPERIENCE
AN APPOINTMENT ORDER OUTPATIENT SCHEDULING SYSTEM THAT IMPROVES OUTPATIENT EXPERIENCE Yu-L Huang Industral Engneerng Department New Mexco State Unversty Las Cruces, New Mexco 88003, U.S.A. Abstract Patent
More informationSecure Cloud Storage Service with An Efficient DOKS Protocol
Secure Cloud Storage Servce wth An Effcent DOKS Protocol ZhengTao Jang Councaton Unversty of Chna z.t.ang@163.co Abstract Storage servces based on publc clouds provde custoers wth elastc storage and on-deand
More informationSTATE HIGHWAY ADMINISTRATION RESEARCH REPORT ENHANCEMENT OF FREEWAY INCIDENT TRAFFIC MANAGEMENT AND RESULTING BENEFITS
MD-11- SP009B4Q STATE HIGHWAY ADMINISTRATION RESEARCH REPORT ENHANCEMENT OF FREEWAY INCIDENT TRAFFIC MANAGEMENT AND RESULTING BENEFITS WOON KIM AND MARK FRANZ GANG-LEN CHANG DEPARTMENT OF CIVIL AND ENVIRONMENTAL
More informationHOUSEHOLDS DEBT BURDEN: AN ANALYSIS BASED ON MICROECONOMIC DATA*
HOUSEHOLDS DEBT BURDEN: AN ANALYSIS BASED ON MICROECONOMIC DATA* Luísa Farnha** 1. INTRODUCTION The rapd growth n Portuguese households ndebtedness n the past few years ncreased the concerns that debt
More informationAssessing Student Learning Through Keyword Density Analysis of Online Class Messages
Assessng Student Learnng Through Keyword Densty Analyss of Onlne Class Messages Xn Chen New Jersey Insttute of Technology xc7@njt.edu Brook Wu New Jersey Insttute of Technology wu@njt.edu ABSTRACT Ths
More informationSoftware project management with GAs
Informaton Scences 177 (27) 238 241 www.elsever.com/locate/ns Software project management wth GAs Enrque Alba *, J. Francsco Chcano Unversty of Málaga, Grupo GISUM, Departamento de Lenguajes y Cencas de
More informationControl Charts with Supplementary Runs Rules for Monitoring Bivariate Processes
Control Charts wth Supplementary Runs Rules for Montorng varate Processes Marcela. G. Machado *, ntono F.. Costa * * Producton Department, Sao Paulo State Unversty, Campus of Guaratnguetá, 56-4 Guaratnguetá,
More informationNEURO-FUZZY INFERENCE SYSTEM FOR E-COMMERCE WEBSITE EVALUATION
NEURO-FUZZY INFERENE SYSTEM FOR E-OMMERE WEBSITE EVALUATION Huan Lu, School of Software, Harbn Unversty of Scence and Technology, Harbn, hna Faculty of Appled Mathematcs and omputer Scence, Belarusan State
More informationContext-aware Mobile Recommendation System Based on Context History
TELKOMNIKA Indonesan Journal of Electrcal Engneerng Vol.12, No.4, Aprl 2014, pp. 3158 ~ 3167 DOI: http://dx.do.org/10.11591/telkomnka.v124.4786 3158 Context-aware Moble Recommendaton System Based on Context
More informationEnterprise Master Patient Index
Enterprse Master Patent Index Healthcare data are captured n many dfferent settngs such as hosptals, clncs, labs, and physcan offces. Accordng to a report by the CDC, patents n the Unted States made an
More informationThe CASIA Statistical Machine Translation System for IWSLT 2008
The CASIA Statstcal Machne Translaton Syste for IWSLT 2008 Yanqng He, Jajun Zhang, Maox L, Lcheng Fang, Yufeng Chen, Yu Zhou and Chengqng Zong Natonal Laboratory of Pattern Recognton, Insttute of Autoaton
More informationTHE DISTRIBUTION OF LOAN PORTFOLIO VALUE * Oldrich Alfons Vasicek
HE DISRIBUION OF LOAN PORFOLIO VALUE * Oldrch Alfons Vascek he amount of captal necessary to support a portfolo of debt securtes depends on the probablty dstrbuton of the portfolo loss. Consder a portfolo
More informationModeling and Simulation of Multi-Agent System of China's Real Estate Market Based on Bayesian Network Decision-Making
Int. J. on Recent Trends n Engneerng and Technology, Vol. 11, No. 1, July 2014 Modelng and Smulaton of Mult-Agent System of Chna's Real Estate Market Based on Bayesan Network Decson-Makng Yang Shen, Shan
More informationRisk Model of Long-Term Production Scheduling in Open Pit Gold Mining
Rsk Model of Long-Term Producton Schedulng n Open Pt Gold Mnng R Halatchev 1 and P Lever 2 ABSTRACT Open pt gold mnng s an mportant sector of the Australan mnng ndustry. It uses large amounts of nvestments,
More informationLeast Squares Fitting of Data
Least Squares Fttng of Data Davd Eberly Geoetrc Tools, LLC http://www.geoetrctools.co/ Copyrght c 1998-2016. All Rghts Reserved. Created: July 15, 1999 Last Modfed: January 5, 2015 Contents 1 Lnear Fttng
More informationLuby s Alg. for Maximal Independent Sets using Pairwise Independence
Lecture Notes for Randomzed Algorthms Luby s Alg. for Maxmal Independent Sets usng Parwse Independence Last Updated by Erc Vgoda on February, 006 8. Maxmal Independent Sets For a graph G = (V, E), an ndependent
More informationPRIVATE SCHOOL CHOICE: THE EFFECTS OF RELIGIOUS AFFILIATION AND PARTICIPATION
PRIVATE SCHOOL CHOICE: THE EFFECTS OF RELIIOUS AFFILIATION AND PARTICIPATION Danny Cohen-Zada Department of Economcs, Ben-uron Unversty, Beer-Sheva 84105, Israel Wllam Sander Department of Economcs, DePaul
More informationA Cryptographic Key Binding Method Based on Fingerprint Features and the Threshold Scheme
A Cryptographc Key ndng Method Based on Fngerprnt Features and the Threshold Schee 1 Ln You, 2 Guowe Zhang, 3 Fan Zhang 1,3 College of Councaton Engneerng, Hangzhou Danz Unv., Hangzhou 310018, Chna, ryouln@gal.co
More informationPSYCHOLOGICAL RESEARCH (PYC 304-C) Lecture 12
14 The Ch-squared dstrbuton PSYCHOLOGICAL RESEARCH (PYC 304-C) Lecture 1 If a normal varable X, havng mean µ and varance σ, s standardsed, the new varable Z has a mean 0 and varance 1. When ths standardsed
More informationGroup Solvency Optimization Model for Insurance Companies Using Copula Functions
Internatonal Conference on Econocs, Busness and Marketng Manageent IPEDR vol.9 () () IACSIT Press, Sngapore Group Solvency Optzaton Model for Insurance Copanes Usng Copula Functons Masayasu Kanno + Faculty
More informationOverview of monitoring and evaluation
540 Toolkt to Combat Traffckng n Persons Tool 10.1 Overvew of montorng and evaluaton Overvew Ths tool brefly descrbes both montorng and evaluaton, and the dstncton between the two. What s montorng? Montorng
More informationEfficient Project Portfolio as a tool for Enterprise Risk Management
Effcent Proect Portfolo as a tool for Enterprse Rsk Management Valentn O. Nkonov Ural State Techncal Unversty Growth Traectory Consultng Company January 5, 27 Effcent Proect Portfolo as a tool for Enterprse
More informationProbabilistic Latent Semantic User Segmentation for Behavioral Targeted Advertising*
Probablstc Latent Semantc User Segmentaton for Behavoral Targeted Advertsng* Xaohu Wu 1,2, Jun Yan 2, Nng Lu 2, Shucheng Yan 3, Yng Chen 1, Zheng Chen 2 1 Department of Computer Scence Bejng Insttute of
More informationThe Packing Server for Real-Time Scheduling of MapReduce Workflows
The Packng Server for Real-Te Schedulng of MapReduce Workflows Shen L, Shaohan Hu, Tarek Abdelzaher Unversty of Illnos at Urbana-Chapagn {shenl3, shu7, zaher}@llnos.edu Abstract Ths paper develops new
More informationThe Application of Fractional Brownian Motion in Option Pricing
Vol. 0, No. (05), pp. 73-8 http://dx.do.org/0.457/jmue.05.0..6 The Applcaton of Fractonal Brownan Moton n Opton Prcng Qng-xn Zhou School of Basc Scence,arbn Unversty of Commerce,arbn zhouqngxn98@6.com
More information