l.INTRODUCTION For example, multiple linear MISSING DATA
Transcription
l.INTRODUCTION For example, multiple linear MISSING DATA
lh International Quality Conference May 24'n 2013 Center for Quality, Faculty of Engineering, University of Kragujevac MISSING DATA ESTIMATION IN D.4TM STRUCTURES USING MULTIPLE IMPUTATION METHOD Aleksandar Novakovicr) Vesna Rankovicr) Dejan Divac2) Nenad Grujovicr) Nikola Milivojevic2) I) Abstract: The ffictive dam safety monitoring prog'ams ,s essential for dqm and rs widely accepted' Instrumentation ss part of dam safety program's is installed to measure a particular parameter of interest. These parameters might include water levels, seepqge flows, deformations or displacements, pressures, loading conditions, temperature variations, seepage water clarity, piezometric levels, etc. The aim o1" the Departmentfor Applied Mechanics and Aut omatic Control, Faculty of Engineer ing, Univ ersity of Kragujevac, Serbia timely detection of abnormal behaviour of the dam does 2) necessarily imply frequent monitoring or the collection of a great deal of data. h is important that this information is representative and adequ,ztely interpreted. Interpretation ofthe availqble data is very substqntial for dam health monitoring. The dstq interpretation csn be dfficult when data are missing or incomplete. In this paper nultiple imputation method not Institutefor Development of Water Resources "Jarosl w Cerni", Belgrade, Serbia was used to estimate replacement values for the mis:sing data. The results of simulotion show that the mulltiple linear regression model for prediction of the water trevel in piezometers with estimated missing values prcwide better results. Keywords: dam, missing data, multiple imputtttion method, piezometric w ater level as the amount of missing data irLcreases, and if the missing data is not disitributed l.INTRODUCTION Most learning algorithms generally completely randomly, can result in assume that training and test datasets are complete. However, real data sets are often severely biased models [1-2]. incomplete and they contain a proportion of missing values due to various reasons affected such as equipment elrors, manual data entry procedures, and incorrect measurements. For example, multiple linear are unable to directly handle missing data. Many regression techniques software implementations of multiple linear regression ommits all instances with missing data before the model is constructed. Such an approach may lead to significant loss of informations, especially 7ft Prediction performances are not if there is less than lVo missing instances, although l%-5% is manageable. However, sophisticated handling rnLethod is 5%-15% missing there required greater than l5o/o missing instances, while data can severely degrade the prediction performance of learning algorithmsr [3]. In response to these issues various solutions have been developed in statistics if is [a-5] and data mining [6-7]. The ffeatment of missing v'alues is determined by the type of missing data' There are three types of missing data, as IQc May,242013 411 fl,,-. ffiwfl*w"*tw$$*,ggm$4,w*x$$W 4'g:$$$*ge$H*"* follows [8]: a) Missing completety at random (MCAR) There is no dependency between missing value for an afibute and any other observed data or missing attribute; b) Missing at random (MAR) - The missing value for an atribute depends on other known data; c) Missing not at random (MNAR) - The missing value for an attribute depends on other missing values, and thus missing data cannot be estimated from observed data. The objective of this study is p, the squares ofthe errors: (",- z.r)' +(2,- z^r)' +...+(r, "*)' Q) in which 2., denotes the MLR output s= value from the i-th input element: z* Bru*, = Fo + Bru.,, + flur, +...+ The matrix form of Eq. (2) is: where: f: to predict the piezometric water level in dam. Two models were compared, one with and one without estimated missing values in their dataset. In this paper it is MAR, which implies that the missing values are deductible in some complex manner from the remaining data. In order to estimate replacement values for the missing data, multiple imputation method u, o =lt. l: (4) uzr : : lt il,, uro P={fof, (3) (z-af) e:(z-up)' to develop a multiple linear regression model assumed that missing values appear to be by which the sum of can be estimated .-. "*,f : : I urrl Fr}', ,={r, r, and the least squares estimator -.- zr\', of 1f is given by: / n r-l f =la'U) U' z (5) was used. 3. MTTLTIPLE 2. REGRESSION ANALYSIS a Regression analysis is statistical technique for investigatrng and modeling the relationship between variables [9]. The multiple linear regression model is widely used for data analysis or prediction in dam engineering [10]. MLR is used for modelling the linear relationship between a dependent variable and one or more independent variables. Consider {(u, a training data r,),{ur, where4 rr),. -.,(u o, z o)} e [ ={u,ur,...rr,\' is a set n "" vector of input variables and z, is the corresponding output value, p is the number of training data points. The multiple linear regression model is given by: z* = 0o+ pp, + prur+...+ B*u, IMPUTATIOI{ METIIOD (l) where B, represents unknown parameters, Multiple imputation is sttatistical approach to the analysis of incomplete data, and in this section its main features are summarized. Detailed description is given in the literature [1 l]. p data matrixo which can be thought of as Lets assume that X is the X = (Xo6o,X*r"), where Xo6, n>< axtd X*;, are the observed and the missing parts, respectively. is considered ar model It P(Xl?) for the data vector parameter. In X, where 0 is a M the beginning, complete datasets are created, using an apprropriate imputation model to generate a prlausible values for the missing observatiions. In order to obtain the imputed valups data augmentation, [12], is used. Practiically, it is a MCMC (Markov Chain Montr: Carlo) procedure in which, given the values d(fr) 4t2 A, Novakovic, V. Rankovic, D, Divac, N. Grujovic, N. Milivojevic $ wx$ww se+e$$ t#) at the fr -th iteration, these values are updated by drawing random and values from the conditional distributions follows: xX:', D P(x^,1x"^,e@) qx*m $ q;sem$ $ and testing $'s { c :y$$q}$ # $$** $]'-&$i$ S.t S$ L#.* I' MLR models. as (6) n P(olx"^,x*:t\ (i) (6) Step is called the Imputation step, g(r+r) while (7) is known as the Parameter step. When # -+ oo, the sequence (t o, ,r*]) has a stationary distribution whose marginals are P(el Xrbr) and P(X^ul Xotr), respectively. After convergence, the imputations are acquired from (6). Finally lets *$) =(X"b".XH), i =1,...M denote the imputed-data estimates Under general condition, imputation estimate follows x M I l]: of X of X. multiple is calculated as (8) 4. CASE STUDY: PRVOT\-EK DAM The dam and the reservoir Prvonek (Fig. l) were built in 2005, in order to solve the water supply problem of the towns Vranje, Bujanovac, and the surrounding villages in south-east Serbia. They are located on the Vranjsko-Banjska River, the right tributary of the river Juhta km upstream of Vranjska Moravq 9 5. SIMI]LATION RESI]LTS For the purpose of construcrting the MLR model, a program was written in R by the authors. The program impk:mented classes provided by the Amelia II package, which offers a comprehensive rimge of functions, necessaxy for implementation of multiple imputation method. Accuracy of the MLR model depends =l?t't Banja-Spa, near Figure 1. The view of Prvonek dum the village Prvonek. Prvonek dam is rockfill embankment dam, with sloped central clay core within the dam body. The height of the dam is 90 m. At maximum water levels, the volume of the reservoir is 20 million m'. For the purpose of this paper, one of piezometers, installed on the section of the the dam, was observed. The water level in examined piezometer have been measured every day. The data collected from June 2010 to April 20ll were used for training 7tr IQC on the appropriate choice of the input variables. The input variables of both MLR models were measurementsl of the tailwater levels taken on the same day (hlr), I day before (hlz) and 2 days before (hl3) the measurements taken by piezometers. For the purpose of training ancl testing MLR models, respectively, T0o/o and 30Yo of randomly chosen data points collected during the period of June 20l0-April 201 I were used. Collected dataset contained 24% of missing data. The MLR models for prediction of water level in the examined piezometer, one without and one with estimated missing values in their dataset, are respectively: hpff -- 194.52 + 0.64. hl, + 0.08 - hl, - 0.06. hl3 (e) hpY = 169.50 + 0.68' hl, + 0.07' hl, - 0.05' hl3 (10) The performance of two MLR models was evaluated by comparing the erstimates of the models with experimental d,ata. The performance parameters of the trairring and test sets are presented in Table l. May,242013 413 fl.,j.ffi*$ $m$*s seaa $$s$m $ $-xs *e$ $ S €ts:,wt$ex Tablel. The performance parumeters of MLR modelsfor prediction of water levets in the examined Piezometer hpff 0.93 Test 0.96 hpY Training 0.97 Test 0.99 m se values in their dataset, were developed to predict water level in one of the installed piezometers installed on the sectio,n of the dam. The performance of the fi\,o MLR r Trainins * models were tested using co.nelation coefficients. As it can be easily observed, both models are capable of predicting water levels in piezometersi with reasonable accuracy, although morCel with estimated missing values 5. CONCLUSION gives a in its dataset slightly higher coefficient of correlation values for training and test sets. In this paper, two MLR models, one with and one without estimated missing RDF'ERENCES: lll Allison, P. D. (2001). Missing Data. Thousand oaks, Sage, cA: Sage universilr papers I2l Little, R. J. A., & Rubin, D. B. (2002), Statistical Analysis with Missing Data. Hoboken, NJ: John Wiley and Sons. t3l Acuna, E., Series on Quantitative Applications in the Social Sciences. & Rodriguez, C. (2004). The treatment of missing values and its effe,ct in the classifer qccuracy. In: Banks, D., House, L., McMonis, F. R., Arabie, P., Gaul, W. (Eds.), Classification, Clustering and Data Mining Applications, Springer, 639-6418. t4l Dempster, A. P., Laird, N. M., & Rubin D. B. (1977). Maximum likelihood from incomplete data via the EM algorithm . Journal of the Royal Statistical Society, 39,l-38. t5l W*g, Q., Rao, J. N. K. (2002). Empirical likelihood-based inference under imputation for missing response data. Annals of Statistics, 30,896-924. t6] Lakshminarayffi, K., Harp, S. A., & Samad, T. (1999). Imputation of missing data in industrial databases. Applied Intelligence, I I , 259-27 5. l7l t8l Chan, K., Lee, T. W., Sejnowski, T. J. (2003). Variational Bayesian learning of ICA with missing data. Neural Computation, /5(8), 1991-201l. Shafer, J. (1997). Analysis of Incomplete Multivariate Data.London, UK: Chaprnan and Hall. Montgomery, D. C., & Runger, G. C. (1994). Applied statistics and probability for engineers. John Wiley & Sons, Inc. [0] Rocha, M., Serafim, J., & Silveira, A. (1958). A method of quantitative interpretation of the results obtained in the observation of the dams.R.84, Q.21, In: VI ICOLD crcngress, New York, 2927-60. [l] Schafer, J. L. (1997\. Analysis of Incomplete Multivariate Data. London: Chapman & Hall. [12] Tanner, M. A., & Wong, W. H. (1987). The calculation of posterior distibution by data augmentation (with discussion). Journal of American Statistical Associacion, tl2, 528- l9l 550. Acknowledgment: The part of this research is supported by Minisf,y of Science in Serbia, Grants III4I007 and TR37013. 414 A. Novakovic, V Rankovic, D. Divac, N. Grujovic' N. Milivoievic "q \\* '"' *; "l, :." ffi.1,-ffi $ geag m$.+:u s.c*x $$q' ff $i*s ff gfr ru q,g" ffiB $g m"++* eq, $l s '".\'${*#S.# 4 s' ,*,: .ei-: $ $: T.International auality Conference w@ CONFEREI\CE MANUAL May 24c 2013, Kragujevac Faculty of Engineering, University of Kraguievac ffi15- $ffirug#gqge{g$ $ffi *$ffi, q,s;sr$*+$ tlgeq $} . International Quality Conference Conference manual 7 ISBN: 978 - 86 - 86663 - 94 - B Editors: Dr Slavko Arsovski, full professor Faculty of Engineering, Kragujevac Dr Miodrag Lazic, full professor Faculty of Engineering, Kragujevac Dr Miladin Stefanovic, associate professor Faculty of Engineering, Kragujevac Technical Editor: Snezana Nestic Faculty of Engineering, Kragujevac Publisher: FACULTY OF ENGINEERING 34OOO KRAGUJEVAC Sestre Janjic 6 CENTER FOR QUALITY 34OOO KRAGUJEVAC Sestre Janjic 6 For pablishers: No. of copies: Printing: Prof, dr Miroslav Babic Prof. h Slavko Arsovski 2OO Faculty of Engineering Kragujevac Copyright @Faculty of Engineering University of Kragujevac,2013. C opyright @ C anter for Quality, Kraguj ev ac, 2 0 I 3. Publication of Conference manual and organization of 7. lnternational Quality Conference is supported by: Department of Educarion, Science and Technological Development of Reputhlic Serbia Izdavanje Zbontkaradova, organizovanje i odrZavanje 7. Intemational Quality Conference podrZalo je: Ministarstvo prosvete, nauku i tehnoloikog rawoja Republike Srhiie il 7ft IQC May,24th2ol3 of .i #1:lj* -;,q r;,i E t'\i fr l. :. ti:\'i t:: s s tittr* Prof. dr Slavko Arsovski, Faculty of Engineering, Kragujevac, serbia, President Tadeusz sikora, The Department of euality Management, cracow uni.versity of Economics, Krak6w, Poland Prof. drTadeja Jere Lazanski, University of primorska, Slovenia Prof. dr Milan Perovic, Faculty of Engineering, podgorica, Montenegro Prof. dr Branislav Marjanovic, University of Johanesburg, SAR Prof. dr Goran Futnik, Univerzitet Minho, portugal Prof. dr Biilent Eker, Namik Kemal University, Tekirdag-Turkey 8. Assoc. Prof. Marti Casadesfs, Universitat de Girona, Girona, Spain Prof. Stanislav Karapehovic, University of Alberta, Edmonton, Canada, 10. Assoc. Prof. Iflaki Heras, Universidad del Pais Vasco, San Sebastian, Spain I L Miroslav Badida, Technical University of Kosice, Faculty Engineering, Department of Environmental, Studies and Information Engineering 12. Prof. dr Mirko Sokovic, Fakultet za strojnistvo Ljubljana, Slovenia 13. Prof. dr Ljupco Arsov, Elektrotehnicki fakultet Skoplje, FYR Macedonia 14. Prof. dr Zdravko Krivokapic, Faculty of Mechanical Engineering, Poclgorica, 2. 3. 4. 5. 6. 7. 9. of 15. 16. 17. 18. 19. 20. Montenegro Prof. Dr. Bernhard Miiller, Leibnizlnstitute of Ecological and Regionall Prof. dr Miodrag Lazic, Faculty of Engineering, Kragujevac, Serbia Prof. dr Janko Hodolic, Faculty of Technical Sciences, Novi Sad, Serbia Prof. dr Miladin Stefanovic, Faculty of Engineering, Kragujevac Prof. dr Ayqegiil Akdogan Eker, Yrldrz Technical University, Mechanical Faculty, Beqiktag/istanbul-Turkey Dr. Prasun Das, SQC & OR Division of the Indian Statistical Institute (ISI), Kolkata,India 21. Georgeta Rafl, U.S.A.M.V.B. Timigoara, Romdnia 22.Prof. dr. Petroman Ioan, USAMVB Timisoara, Romania 23. Paul M. Andre, AQE Group, Chicago,Illinois, USA 24. Prof. dr Ezendu Ariwa, London Metropolitan Business School, l-ondon Metropolitan University, UK 25. Paul M. Andre, AQE Group, Chicago,Illinois, USA 26. Nenad Injac, Quality Austria, Wien, Austria 27. dr Kresimir Buntak. Tehnicko veleuciliste Yarazdin. Croatia 7th IQC May,24th2ol3 III $s l. sg*rs'se*$$$ *x {$$ {.s*a* $ $$+' $'$ +e $qr$ $:e}*:* $j, }L{ $ Shirshendu Roy, Samar Bhattacharyay, prasun Das LEARNING IMPACT, LEARNING MODE AND E-LEARNING_ AN INDIAN 2. SCENARIO .....................3 Petroman cornelia, Petroman I., sdrdndan H., Marin Diana, viduva Loredana, $ucan Moisina TMPROVING QUALITY MANAGEMENT rN TTrE BEEF TNDUSTRY .........9 3. Rouhollah Mojtahedzadeh,Rezalzadi APPLYING A THEOR-ETICAL MODEL FOR ORGANIZATIONAL DECISION MAKING BASED ON SCHOOL OF INTELLECTUAL STAFF (MODERNISM, SYMBOLIC, POST MODERNISM)...........................15 4. Rouhollah Mojtahedzadeh, Reza Izadi APPLICATION OF PERFORMANCE APPRAISAL SYSTEM IN DEVELOPING 5. COUNTRIES .........3I Rouhollah Mojtahedzadeh, Reza Izadi LOCAL GOVERNMENT ADMINISTRATION IN SOUTH AFRICA: T}M MYTH, T}IE PARADOX, TI{E CHALENGES AND THE WAy 6. FORWARD .......................37 RouhollahMojtahedzadeh TI{E RULE OF SCHOOL-BASED MANAGEMENT IN DEVELOPING COUNTRIES .......,,47 7. Rouhollah Mojtahedzadeh,Rezalzadi THE IMPACT OF HUMAN RESOURCE MANAGEMENT ON PERFORMANCE OF OIL AND GAS INDUSTRY IN IRAN ...........................57 8. Biilent Eker, Aygegiil Akdogan Eker OPPORTIJNITIES TO USE IMAGE PROCESSING TECHNOLOGY IN QUALITY-BASED PRACTICES ........................67 9, Btlent Eker, Ayqegiil Akdolan Eker MATLAB-BASED APPLICATIONS FOR IMAGE PROCESSING AND rMAGE QUALITY ASSESSMENT .................. ...................73 10. Dragan Cvetkovic, Milorad Bojic,Verimir Stefanovic, Dragan Taranovic, Marko Miletic, Sasa Pavlovic DEVELOPMENT OF EXPERIMENTAL PROCEDURE FOR INVESTIGATION LOW-TEMPERATURE }IEATING SYSTEMS... ..,,...,.......7 9 l. Jasna Glisovic, Jovanka Lukic, Danijela Miloradovic, Dobrivoje Catic RESEARCH OF THE ruSTIFICATION FORTI{E HIGH ..................85 WARRANTY COST DUE TO DISC BRAKE I NOISE 7ft IQc May,24h2ol3 V $ .*,\t $ sq$$ffi *$ $$e, mm $ e;m**$$$p, $ ,*x+ $,seeffi $ :* 12. Nikola Petrovic MODIFIED SERVQUAL MODEL OF SERVICE QUALITY MEASUREMENT IN HOTELS WITH BUSINESS FACILITIES ..................... 9 I 13. Wieslaw tr ukasiriski SELF ASSESSMENT AS A SOURCE OF KNOWLEDGE IN THE PROCESS OF AN ORGANISATION IMPROVEMBNT.... .................9,1, 14. Predrag Pravdic AN INTEGRATED MANAGEMENT SYSTEMS BASED ON STEP 15. STANDARD ..........109 Predrag Pravdic MANAGING BUSINESS GOALS OF MANT]FACTURING ORGANIZATIONS BY 16. BSC....... ...............I19 Predrag Pravdic QUALITY OF CAD/CAM SOFTWARES IN MANUFACTIJRING 17. INDUSTRY .,.,.,.,,...,......,..,127 Srdjan Bogetic, Dejan Djordjevic, Dragan Cockalo ANALYSIS OF THE PROCESS OF PROMOTING CORPORATE SOCIAL RESPONSIBILITY IN FUNCTION oF COMPETITIVENESS IMPROVEMENT.................. ...............143 18. Bojan Stojcetovic, Milan Misic, Zivce Smkocevic QUALITY TOOLS IN PROJECT MANAGEMENT..........................................153 19. Bedri Onur Kucukyildirim, Aysegul Akdof,an Eker QUALITY ASSESSMENT OF CARBON NANOTUBES: CHOOSING CHARACTERZATION METHOD .........................I59 20. Dragan Lazarevic, Milan Eric, Milan Misic THE DEVELOPMENT OF DIGITAL FACTORY IN TODAY'S WORLD......165 21. Boris Agarski, Branislav Milanovic, Darko Milankovic, Milana Ilic, Igor Budak, Djordje Vukelic, Janko Hodolic APPLICATION OF ANALYTIC HIERARCHY PROCESS FOR WEIGHTING OF IMPACT CATEGORIES INLIFECYCLEIMPACTASSESSMENT.................. 22. Jelena Cadjenovic Milovanovic KNOWLEDGE oF MODERN 23. ...,.............181 - INDISPENSABLE RESOURCE COMPANY. ..................189 Jelena Cadjenovic Milovanovic FROM CRM & KM TO CUSTOMERKNOWLEDGE MANAGEMENT ........197 VI 7ft IQC May,2462ol3 $*s$'*l$'$e g$e sxm$ {isxl*#${$ fl+'*x$qr#eru* 24. 25. JelenaCadjenovic Milovanovic GREEN OFFICE _ FUTURE OF NEW AGE 27. TECHNOLOGY............... Alexandre Patou-Parvedy, Milan Despotovic MODELING OF A SOLAR CHIMNEY ON ........................2r3 ENERGYPLUS SERBIA ..,......231 Jovan Malesevic, Milos Milovanovic, Slobodan Djordjevic, Milorad Bojic, Nebojsa Lukic TTM INFLTIENCE OF THE TROMBE WALL ON ENERGY CONSUMPTION FOR HEATING AND COOLING "NET ZERO'' HOUSE 29. 221 Ielena Cadjenovic Milovanovic, Maja Angelovski SOME OF THE I]DEAS HOW TO IMPROVE YOUTH ENTREPRENEURSHIP IN 28. ....................205 Danijela Nikolic, Milorad Bojic, Jasmina Skerlic, Jasna Radulovic, Marko Miletic A REVIEW OF SILICON SOLAR CELLS rN pHoTovoLTArCS 26. ............ * $l'$t$; ...,...237 Nenad Kostic, Mirko Blagojevic, Vesna Marjanovic, Zorica Djordevic, Milorad Bojic DETERMINATICN OF SOLARANGLES FOR SUITABLE POSNIONING OF SOLAR SYSTEMS FOR PARTICULART]MES OF THEYEAR.......... 30. Milos ,..,....,.24,9 Matejic, Mirko Blagoj evic, ZoricaDjordjevic, Nenad Marjanovic,, Nenad Pefiovic COMPARATIVE ANALYSIS OF DIFERENT TYPE REDUCERS FOR WTNCH DRIJM DRIVING UNIT........... 31. ..,,.....,25.5 Bernard Bincrycki ERGONOMICS N{ HUMAN RESORCES MANAGEMENT....... ..,.,....,.......,...26.J 32. Pawel Nowicki RISK MANAGEIVIENT AN IMPORTANT ISSUE IN QUALITY MA.NAGEMENT 33. SYSTEMS Piotr Kafel FOOD QUALITY PRODUCTS IN EU 34. Nikola Rakic, Milam COUNTRIES .............261 ...................27:3 Popovic, Dusan Canovic, Nebojsa Jovicic, Ivlilun Babic ENVIROMANTAL AND FINANCIAL ASPECTS OF REPLACING COI\L AND FUEL OIL WITHNATURAL GAS ON THE "HOME LOCATION'' OF "ENERGY'LTD..... ze Iqc May,z4th2ol3 ...,....,....271' VII .{.j-19$ $ss$*ry** 35. 36. $e+,*ew$ g**+s$$ffi *.sffisss,s Novak Nikolic, N,ebojsa Lukic, Dragan Taranovic MEASUREMEN'T CHAIN FOR THE BIFACIAL AND TI{E CONVENT.TONAL FLAT-PLATE WATER SOLAR COLLECTORS...... LIGHTING.............. .............,.........29t1 Nenad Petrovic, Ivtirko Blagojevic, Nenad Marjanovic, Milos Matejic PARAMETRIC DRAWING OF A CYCLO DRIVE SHORTENED E()UIDISTANT EPITROCHOID GEAR.. 38. 2 Ti5 Slobodan Djorde'i,ic, Milorad Bojic, Dragan Cvetkovic, Jovan Malesevic, Marko Miletic INFLUENCE OF HOUSE SHADOWING TO TTIE CONSUMP'NON OF PRIMARY ENERGY FOR ITEATING, COO.LING, AND 37. $ .............303 Hajnalka Kovac Sarkanj, Vilmos Kovac FOOD SAFETY I{ND QUALITY MANAGEMENT SYSTEMS APPLIED IN SEF|.BIAN AND HUNGARIAN FOOD CHAINS ........................ 309 39. Zeljko Spijunovic, Radosav Mirkovic, Zoran Nesic, Miroslav Radojicic, Jasmina Vesic Vasovic AUTOMATIZATION OF OBTAINING INFORMATION AT INSPECTION OF'TECHNICAL SAFETY OF RAILWAY VEHICLES ...........3I5 40. Sasa Bogicevic, Jasmina Vesic Vasovic, Miroslav Radojicic, Zoran Nesic APPLICATION OF CLUSTER ANALYSIS IN FUNCTION OF IMPROVING DECISION MAKING PROCESS ...,.................325 41, ZorcnNesic, Jasmina Vesic Vasovic, Miroslav Radojicic IMPROVEMENT OF TIME ANALYSIS QUALITY IN NETWORK DIAGRAMS BY IMPLEMENTATION OF SOFTWARE SUPPORT ........,,.....329 42. Danijela Tadic, Jo'vana Kostic, Marija Zahar Djordjevic, Hrvoje Puskaric TFIE PLANT WASTE MANAGEMENT PROBLEM IN UNCERTAIN EWIRONMENT......... 43. .................335 Jasmina Skerlic, IV[ilorad Bojic, Danijela Nikolic, Jasna Radulovic, Ivlarko Miletic A REVIEWLIFE CYCLE ASSESSMENT OF A SOLAR THER]\4AL COLLECTOR SENSITIVITY ANALYSIS, ENERGY AND ENVIRONMENTAL 44. ZoricaDjordjevic, Sasa BALANCES ......................34I Jovanovic, Milorad Bojic, Dragan Adamovic', Milos Matejic THE INFLUENC]E OF TV AND VIDEO APPLIANCES AND INFORMATION TECHNOLOGY EQUIPMENT ON ENERGY COINSUMPTION IN HOUSEHOLDS............... .....351 VUI 7s IQC May,z4th2oL3 " $gx$*x,m*$$$,* 45. g,#$: $j$$ffi #$€+,, 4. q r*#$ "$$ 1 *:*$+ r:,* Jasna Radulovic, lMilorad Bojic, Danijela Nikolic, Jasmina Skerlic, Dragan Taranovic TOWARDS NET ZERO ENERGY BUILDINGS: POSSIBILITIES FORPHOTOVOLTAIC 46. USE ........3:i7 Jovanovic, IVlarko Miletic, ZoicaDjordjevic, Ivan Miletic, Miknad Bojic OPTIMISATION OF ZERO-NET ENERGY HOUSE ORIENTATION rN CTTIES OF DTFFERENT ....................363 Sasa LATrTUDE.................. 47. Dragana Rejman Itetrovic, Zora Arsovski, Vladimir Rankovic, Zoran Kalinic, Igor Milanovic BUSINESS PROCESSES MAPPING IN E-SUPPLY CHAIN 48. ....,.369 Fkvoje Puskaric, \fiarijaZahar Djordjevic DETERMINATION OF A DEVELOPMENT PROCESS PERFORMANCE; USING DEVELOPED FU ZZY EXPERT SYSTEM..... ....,,,,37 49. Aleksandar Vujovic, Zdravko Krivokapic, Jelena Jovanovic, Sanja Pekovic, Radivoje Micunovic BUSINESS PROCESS IMPROVEMENT BY APPLYING EIENCHMARKING BASED 50. 5 MODEL ................385 Piotr Kafel, Jelena Jovanovic, Zdravko Krivokapic, Aleksanadm Vujovic MONITORING, IdEASUREMENT, ANALYSIS AND REVIEW IN POLISH AND MONTENEGRIN ORGANIZATION ACCORDING TO THE ISO 9OO,{ MANAGEMENT MATURITY MODEL.............................395 51. Rouhollah Mojtahr:dzadeh ROLE OF BOAR]D OF BOARD SIZE ON FINANCIAL PERFORMANCE OF COMMERCIAL BANKS ...................405 52. Aleksandar Novakovic, Vesna Rankovic, Dejan Divac, Nenad Grujovic, llikola Milivojevic MISSING DATA ESTIMATION IN DAM STRUCTURES USING MULTIPI,E IMPUTATION METHOD .....,.411 53. Milan Despotovic, Sasa Babic, Jovanka Lukic, Branimir Milosavljevic APPLICATION OF ARTIFICIAL NEURAL NETWORK FOR PREDICTION OF TRAFFIC NOISE BASED ON THE TRAFFIC FLOW STRUCTURE ................4I5 54. Aleksandar Jankulovic, Miroslava Raspopovic PERFORMANCE MEASURES ANALYSIS FOR ONLINE AND TRADITIO]\IAL 55. STUDIES ......,..423 Goran Manojlovic. Slavko Arsovski, Ivica Nikolic INTRODUCTTO\I OF QUALrry STANDARDS FOR SCHOOLS... ................ 42'9 7fr IQC May,24h2ol3 x fl." ,$,t]$ $eetsi x *$e* $$* se+e$ $jm*$$$.+ $. r,s+ #es *sere, 56. Lozica Ivanovic, lDanica Josifovic, Andreja Ilic, Vukic Lazic ECOLOGYCAL ASPECTS OF HIGH STRENGTH LOW ALLOYED STEELS AT MECHANICAL CONSTRUCTIONS ..............435 57. Lozica lvanovic, ,\ndreja Ilic, Blaza Stojanovic,Katarina Zivkovic TFIE INFLUENCE OF DESIGN MODIFICATIONS OF CARDAN SHAFT DRIVEN FORK ON ITS STRESS DISTRIBUTION............... .....,4111 58. Aleksandar Maric, Zoranapavlovic, Slavko Arsovski IMPROVING TIIIE QUALITY OF ORGANIZATIONAL AND FUNCTIONAL CONCEPTS OF PUBLIC ADMINISTRATION BY INTRODUCING PROJECT FLINCTION............... .,...............4417 59. Snezana Nestic, Nliladin Stefanovic, Aleksandar Djordjevic, Slavko Arsovski, Svetlana Stojanovic AN ASSESSME].{T AND OPTIMIZATION oF QUALITY oF STRATEGY 1PROCESS.................. 60. .....................453 TudorPendiuc INDICATORS AI\ID IMPLICATIONS FOR INFORMATION QUALITY MANAGEMENT rMPROVEMENTS............. .....46 5 61. Aleksandar Djordjevic, Milan Eric, Aleksandar Aleksic, Snezana Nestic, Svetlana Stojanovic OPTIMIZATION OF MACHINING PROCESSES USING THE ABC METH.OD AND GENETIC 62. Snezana Pesic-Djokic, Ivan ALGORITHM............ ..,...471 Djokic QUALITYAND'WORLD CLASS MANUFACTURrNG..................................483 63. Srdjan Nikezic, Dobrica Stojkovic, Boban Djurovic, Aleksandar Djordjevic LEADERSHIP NIETWORK BLAKE, MOUTON AND MCCANSE: CASE STUDY - I.EADERSHIP STYLES AND DIMENSIONS IN ONE OF TTIE LOCAL SELF-GOVERNMENTS IN SERBIA....,.................489 64. Goran Boskovic,l{ebojsa Jovicic, Marko Milasinovic, Gordana Jovicic, Dobrica Milovanovic METHODOLOG'T FORREDUCTION OF GHG EMISSIONS FROM MUNICIPAL SOLID WASTE COLLECTION AND TRANSPORT.....5O5 65. Vesna Rankovic, Irlilorad Bojic, Aleksandar Novakovic, Dragan Cvetkovic, Marko Miletic FU ZZY CONTROLLER SYNTHESIS FORBUILDING SHADING CONTROL .................517 66. Milorad Bojic, Dragan Adamovic, Jasna Radulovic, Marko Miletic, Ljrubisa Bojic LIGHTING USE IN SERBIAN LOW-RISE HOUSES..... X zft Iqc May,24h2ol3 .............623 $sx$+w.m*$$+. 67' * a*g$ $.,gses*$$ g 4 € $ e${ric +lffs* u :{,;q{$ ve.sna Radonjic-Djogatovic, Aleksandra Kostic-Ljubisavljevic, Mrjana Stojanol,ic QUALITY OF BUSINESS CONSIDERATIONS IN TELECOMMUNICATIONNETWORKS 68, Y:!nl ...,....,..5,.2g Radonjic-Djogatovic, Aleksandra Kostic-Ljubisavljevic, Mirjana Stojanovic QUALITY oF E-XPERIENCE MEASUREwNis IN TELECOMMI.INICATIONNETWORKS ...........535 69. Milorad Bojic, Dragan Adamovic, Jasna Radulovic. Ivan Miletic, Ljubisa Bojic AIR CONDITIONING IN SERBIAN LOW-RISE HOUSES 70. Marko Miletic, Milorad Bojic, Ivan Miletic, Nenad Miloradovic, Jasmina skerlic WINDOWS SEL]ECTION INFLUENCE ON ENERGY HEAT GAIN AND Loss IN 71. ........5It1 HousE ........................s4t7 slavko Arsovski, lMiodrag Lazic, srecko curcic, sandra Milunovic MOBILE PRESS FOR CAR RECYCLING DEVELOPMENT..........................553 72. MilanPavlovic, Sr,ecko curcic, Aleksandar Tomovic, Sandra Milunovic NATIONAL ECCNOMY RESOURCE CAPABILITY FOR PRODUCTION R3CYCLING EQUIPMENT FOR MOTOR VEHICLES ........559 73. Igor Milanovic, Miladin Stefanovic, Zora Arsovski, Dragana Rejman Petrovic, Vladimir Rankovic, Zoran Kalinic SUPPLY CHAIN INFORMATION INTEGRATION THROUGH SERVICE ORTENTED ARCHITECTURE 74. ..............,567 Pavle Mijovic, Evzrnthia Giagloglou, Branislav Jeremic, Ivan Macuzic, Marko Djapan, Marko Milosevic INFLUENCE OF PROCESSING ON COSMETIC, PHARMACEUTICAL AND FOOD EMLILSIONS QUALITY, STABTLITY AND RHEOLOGY ........s79 75. Marko Milosevic, lvan Macuzic, Petar Todorovic, Marko Djapan, Evanthia Giagloglou, Djordje Vuckovic IMPLENMNTATION OF THE 55 SYSTEM AS A FACTOR FOR TMPROVTNG THE QUALTTY MANAGEMENT......................................585 76. Katartna Kanjevac Milovanovic THE REVIEW OI' PROBLEM AND THE ADVANTAGE OF INVESTING INA PROJECT OBTAINING CE MARK .....,..59I 77. Milan Radenkovic,, Branislav Jeremic, Petar Todorovic, Marko Djapan, Marko Milosevic, Pavle Mijovic IMPROVEMENT OF QUALITY IN PRODUCTION PROCESS BY APPLYING KAIKAKU METHOD 7tr IQC May,z4thzol3 ....................60I XI 78. Aleksandar Aleksic, Marko Djapan, Miladin Stefanovic, Slavko Arsovski, Danijela Tadic, Ivan Macuzic ANP AS A TOO]L FORDETERMINING THE RESILIENCE F/TCTORS'INTERACTION IN 79. SMES.......... ........607 srdjan Nikezic, D'obrica Stojkovic, Boban Djurovic, Aleksandar Djordjevic LEADERSHIP TRAITS: RESEARCH oF LEADERSHIP AND COLEADERSHIP PERSONAL CHARACTERISTICS IN LOCAL GOVERNMENTS OF SERBIA AS A FRAMEWORK I'OR SOCIAL CHANGE.... .............613 80. Dobrivoje Catic, lvlilorad Bojic, Jasna Glisovic, Zorica Dj ordjevic, Nada Ratkovic FAULT TREE A]VALYSIS OF SOLAR CONCENTRATORS.......... .......,........62:"9 81. Evanthia Giagloglou TOXIC MIXTURES A PROBLEMATIC CASE OF TOXICTTY QUALITY INDEXES.... .........................639 rJQR.......... .........64s 82. Rouhollah Mojtahedzadeh, Reza lzadi THE EFFECT OF'COMPETITIVE ADVANTAGE AS AN INTERVENING'VARI,ABLE BETWEEN CRITICAL SUCCESS FACTORS OF SUPPLY CHI\IN MANAGEMENT IMPLEMENTATION AND PERFORMANCE; IN THE IRANIAN AUTOMOBILE INDUSTRY ....,.,........647 83. Rouhollah Mojtahe,dzadeh, Reza Izadi ACHIEVING ORGANIZATIONAL EFFECTIVENESS THROUGH TQM. PRINCIPLES IN DEVELOPING INDUSTRY: A CASE STUDY OF PALM OIL MECHANDISING BUSINESS rN cRoss RrvElt STATE........ 84. ElizabetaMitreva THE SUPERIOR CUSTOMER'S VALUE OF THE NEW ECONOMY IMPLEMENTED W]THIN MACEDONTAN COMPANrES.............. 85. Vanya Zhivkova CURRENT TRE}IDS IN THE USE OF ...........6s7 ...................669 SOLARENERGY .,.,,,,.,.675 86. TuRunsheng BEHAVIOR QUA.LITY DECIDES OUR CONDITIONS OF SURVIVAL AND DEVELOPMENT.................. .....................683 87. Karin Kandananond THE APPLICATION OF BOX-BEHNKEN METHOD TO OPTIMIZE T]IE DESIGN OF EWMA CHART FOR AUTOCORITELATED XII PROCESSES................ 7tr Iec May,24th2ol3 ....................691 $+r$qi$"liifi{+ +**$ 88. Shahab Alam { tt,ufilu, { 'l ggil'u**u* *, if',fit,i Malik MEASURING S]ERVICE QUALITY PERCEPTIONS OF THE CUSTOIMERS oF RESTAURANTS IN PAKISTAN..........................7I)5 89. SantoshKadam WHYMANAGEMENT SYSTEMFAILS zfr Iqc May,z4h2ot3 ................71g XIII CIP - Haraaoruaaquja y nytSnuxa4nlu Hapo4na 6n6nnorexa Cptiuje, Eeorpa.q 00s.6(082) INTERNATIONAL euality Conference (7 ;2013 ; Kragujevac) Conference Manual / 7. lnternational Quality Conference, May 24rd 2013, Kragujevac ; [organized by] Faculty of Engineering University of Kragujevac ; [editors Slavko Arsovski, Miodrag Lazic, Miladin Stefanovicl. - Kragujevac : Faculty of Engineerin g, ZO1r3 (Kragujevac : Maiinac). - Xil, 706 str. : ilustr. ; 24 cm Tekst Stampan dvostubacno. - Tirai 100. Str. lV: [Preface] / Slavko Arsovski. Napomene i bibliografske reference uz tekst. - Bibliografija uz svaki rad. rsBN 978-86-86663-94-8 1. Faculty of Engineering (Kragujevac) a) MeHapnneHT rorarHilM KBarrreroM - 36opxr,rqra coBtss.sR-tD 198330635