Amit Choudhary* Deptt. of Comp. Sc. Maharaja Surajmal Institute, New Delhi, India. Rahul Rishi
|
|
- Berniece Gordon
- 7 years ago
- Views:
Transcription
1 Amt Choudhary et al/(ijcse) Internatonal Journal on Computer Scence and Engneerng, Optmal Feed Forward MLPArchtecture for Off-Lne Cursve Numeral Recognton Amt Choudhary* Deptt. of Comp. Sc. Maharaja Surajmal Insttute, New Delh, Inda. Rahul Rsh Deptt. of Comp. Sc.and Engg. TITS, Bhwan, Haryana, Inda. Savta Ahlawat Deptt. of Comp. Sc.and Engg. Maharaja Surajmal Insttute of Technology, New Delh, Inda. Vaypal Sngh Dhaa Deptt. of Comp. Sc. IMS, Noda, UP, Inda. Abstract The purpose of ths wor s to analyze the performance of bac-propagaton feed-forward algorthm usng varous dfferent actvaton functons for the neurons of hdden and output layer and varyng the number of neurons n the hdden layer. For sample creaton, 250 numerals were gathered form 35 people of dfferent ages ncludng male and female. After bnarzaton, these numerals were clubbed together to form tranng patterns for the neural networ. Networ was traned to learn ts behavor by adjustng the connecton strengths at every teraton. The conjugate gradent descent of each presented tranng pattern was calculated to dentfy the mnma on the error surface for each tranng pattern. Experments were performed by selectng dfferent combnatons of two actvaton functons out of the three actvaton functons logsg, tansg and pureln for the neurons of the hdden and output layers and the results revealed that as the number of neurons n the hdden layer s ncreased, the networ gets traned n small number of epochs and the percentage recognton accuracy of the neural networ was observed to ncrease up to certan level and then t starts decreasng when number of hdden neurons exceeds a certan level. Keywords- Numeral Recognton, MLP, Hdden Layers, Bacpropagaton, Conjugate Gradent Descent, Actvaton Functons. I. INTRODUCTION The process of offlne numerals recognton nvolves the automatc converson of numeral s gray scale/ bnary mages obtaned from paper documents, photographs etc. nto dgt codes that are nterpreted by the computer and other textprocessng applcatons. The OCR technology has mportance n varous felds le Banng System, Postal Department etc to recognze courtesy amount on ban checs [2] and to recognze postal pn codes wrtten on post cards / letters. The automated processng of handwrtten materal optmzes the processng speed as compared to human processng and shows hgh recognton accuracy and acheves hgh benefts n monetary terms. Recognton of handwrtten numerals s a very complex problem. The dgts could be wrtten n dfferent dmenson, thcness or slant [3]. These wll gve unlmted varatons. The capablty of neural networ to generalze and ts robust nature would be very benefcal n recognzng handwrtten dgts. The artfcal neural networ structure nvolved n our experment of handwrtten numeral recognton s a mult layered perceptron (MLP) wth one hdden layer whch s a supervsed learnng networ n nature. A MLP s a feedforward neural networ that uses error bac-propagaton algorthm for ts tranng [5]. It s a modfcaton of the standard lnear perceptron n that t uses one or more hdden layers of neurons wth non-lnear actvaton functons and s more powerful than the perceptron n that t can classfy data that s not lnearly separable or separable by a hyper plane. The experments conducted n ths paper have shown the effect of the number of neurons n the hdden layer and the transfer functon used n the hdden layer and output layer on the learnng and recognton accuracy of the neural networ. The crteron to judge the performance of the networ wll be the number of tranng epochs, the MSE (Mean Square Error) value and the percentage recognton accuracy. All other expermental condtons such Learnng Rate, Momentum Constant ( ), Maxmum Epochs Allowed, Acceptable Error Level and Termnaton Condton were ept same for all the experments. The rest of the paper s organzed as follows: Secton II deals wth the overall system desgn and the varous steps nvolved n the OCR System [4]. Neural Networ Archtecture s presented n detal n secton III. Secton IV provdes varous expermental condtons for all the experments conducted under ths wor. Functonng of the proposed system s explaned n Secton V. Dscusson of Results and * Correspondng Author
2 Amt Choudhary et al/(ijcse) Internatonal Journal on Computer Scence and Engneerng, nterpretatons are descrbed n secton VI. Secton VII presents the more threshold value. Ths parameter s decded as shown the concluson and also gves the future path for contnual wor n Table : n ths feld. II. OVERALL SYSTEM DESIGN A typcal handwrtten numeral recognton system s characterzed by a number of steps, whch nclude () Dgtzaton / Image Acquston, (2) Preprocessng, (3) Image Bnarzaton (4) Networ Tranng (5) Classfcaton / Recognton and (6) Testng. Fg. llustrates varous steps nvolved n a handwrtten dgt recognton system. Fgure. Typcal Off-Lne Handwrtten Numeral Recognton System A. Preprocessng Preprocessng ams at elmnatng the varablty that s nherent n cursve and hand-prnted numerals. Scalng, Nose Elmnaton, Slant Estmaton and Correcton and Contour Smoothng are some preprocessng technques that have been employed n an attempt to ncrease the performance of the recognton process. Scalng sometmes may be necessary to produce numeral mages of equal sze. Nose (small dots or blobs) may be ntroduced easly nto an mage durng mage acquston; therefore, these small foreground components are usually removed. S. Srhar, V. Govndaraju and R. Srhar also analyzed the sze and shape of connected components n a numeral mage and compared them to a threshold to remove the nose []. Slant estmaton and correcton s an ntegral part of any numeral mage preprocessng. The slope can be estmated through analyss of the slanted vertcal projectons at varous angles. Contour Smoothng [7] s a technque to remove contour nose that s ntroduced n the form of bumps and holes due to the process of slant correcton. TABLE I. INTENSITY / THRESHOLD COMPARISON TABLE Gray-Scale Intensty of the Threshold Text (I) The frst column of Table- represents the ntensty of handwrtten dgt present n the document. Ths ntensty s a gray-scale value when the document s saved after scannng n gray scale mage format and the second column of ths table represents the correspondng value of threshold level for bnarzaton process [6]. The threshold parameter along wth the grayscale mage s made an nput to the bnarzaton program desgned n MATLAB. The output s a bnary mage shown n Fg. 2(a). C. Reshape and Reszng of Numerals for Sample Creaton Each dgt s frst converted nto a bnary form and then reszed to 0x0 matrx M usng nearest neghborhood nterpolaton and reshaped to a 00x logcal vector so that t can be presented as nput to the networ for learnng. The matrx s reshaped [8] usng followng algorthm: Algorthm: ReshapeCharacter(M,00,). Fnd the transpose of Character s bnary matrx M. 2. Read every element vertcally.e. column wse. 3. Save t n the new matrx accordng to ts dmensons n row frst manner. 4. Repeat step 2-3 untl every element s read. B. Bnarzaton of Numeral Images All hand prnted numerals are scanned nto gray scale mages. Each numeral mage s traced vertcally after convertng the gray scale mage nto bnary matrx. Ths bnarzaton s done usng logcal operaton on gray ntensty level as followes: If (I >= Threshold) Else I = 0 I = where 0 < Threshold <= Ths threshold parameter s based on the gray-scale ntensty of the text n the document. More ntensty leads to Fgure 2. (a) Bnary Representaton of Dgt 0 ; (b) Bnary Matrx representaton and (c) Reshaped sample of Dgt 0. The output of ths algorthm s a bnary matrx of sze 00x whch s made as an nput to the neural networ for learnng and testng. Bnary matrx representaton of dgt 0 can be defned as n Fg 2(b). The Reszed dgts were clubbed together n a matrx of sze (00, 0) to form a SAMPLE as shown n Fg. 3. 2
3 Amt Choudhary et al/(ijcse) Internatonal Journal on Computer Scence and Engneerng, So the fnal samples for networ tranng/testng were made n as follows: y f ( Z W ) (3.) Where f s the output functon, Z s the output of hdden layer and W s the weght between hdden and output layer. And, also for hdden layer s processng unt output ; Z n f ( V X ) (3.2) j j j Fgure 3. Input tranng / testng sample preparaton For sample creaton, 250 numerals were gathered form 35 people of dfferent ages ncludng male and female. After the preprocessng s over, 5 samples were consdered for tranng such that each sample was consstng of 0 dgts (0-9) and 3 samples were consdered for testng the recognton accuracy of the networ. III. NEURAL NETWORK S ARCHITECTURE USED IN THE RECOGNITION PROCESS The archtecture selected for the neural networ s a Multlayer Perceptron (MLP) wth one hdden layer. The networ has 00 nput neurons that are equvalent to the nput numeral s sze as we have reszed every numeral nto a bnary matrx of sze 0 X 0. The 0 output neurons correspond to 0 dgts. The number of hdden neurons s drectly proportonal to the system resources. The bgger the number more the resources are requred. The processng nodes of nput layer used the lner actvaton functon and the nodes of hdden and output layers used the non-lner dfferentable actvaton functon as shown n Fg 4. Where X j s the output of nput layer and weght between nput and hdden layer. V s the The MSE between the desred and actual output of the networ s gven by: as; E [ t y ] (3.3) Where t s desred output The error mnmzaton can be shown as; E ' [ t y ] f ( y z W ) (3.4) Weghts modfcatons on the hdden layer can be defned E y z (3.5) ' V *( )[ t y ] f ( y V ) v ' [Let, = t y ] f ( ) ] and we have [ y ' V * W f ( z ) (3.6) Thus the weght updates for output unt can be represented as; W ( t ) W ( t) W ( t) W ( t ) (3.7) Where W (t) s the state of weght matrx at teraton t j W ( t ) s the state of weght matrx at next teraton Fgure 4. Archtecture of the Neural Networ used n the Recognton System. The output of the networ can be determned as: ( t ) W s the state of weght matrx at prevous teraton. and (t) s current change/ modfcaton n weght matrx W 3
4 s standard classcal momentum varable to accelerate learnng process and depends on the learnng rate of the networ. The neural networ was exposed to 5 dfferent samples as acheved n secton III, each sample beng presented 00 tmes. Actual output of the networ was obtaned by COMPET functon. Ths s a compettve transfer functon whch wors as follows: Functon Compet (net N) Start: [S, Q] = Sze (N) (N beng networ smulated vector) [Maxn, ndn] = max (N, [ ], ) Result Matrx A = Sparse (ndn, : Q, Ones (, Q), S, Q) End Functon. Ths COMPET functon puts at the output neuron n whch the maxmum trust s shown and rest neuron s result nto 0 status. The output matrx s a bnary matrx of sze (0, 0). The output s of the sze (0 0) because each dgt has 0 output vector. Frst 0 column stores the frst dgt's recognton output, the followng column wll be for next dgt and so on for 0 dgts. For each dgt the 0 vector wll contan value at only one place. For example dgt 0 f correctly recognzed, wll result n [, 0, 0, 0 all 0]. The dfference between the desred and actual output s calculated for each cycle and the weghts are adjusted by the equaton (3.7). Ths process contnues tll the networ converges to the allowable error or tll the maxmum number of tranng epochs s reached. IV. EXPERIMENTAL CONDITIONS The varous parameters and ther respectve values used n the learnng process of all the experments wth varous actvaton functons of the neurons n hdden and output layers and havng dfferent number of neurons n the hdden layer are shown n Table 2. TABLE II. EXPERIMENTAL CONDITIONS OF THE NEURAL NETWORK Parameters Value Input Layer No. of Input neurons 00 Transfer / Actvaton Functon Lnear Hdden Layer No. of neurons Between 5 and 95 Transfer / Actvaton Functon LogSg or Tansg or Pureln Learnng Rule Momentum Output Layer No. of Output neurons 0 Transfer / Actvaton Functon LogSg or Tansg or Pureln Learnng Rule Momentum Learnng Constant 0.0 Acceptable Error (MSE) 0.00 Momentum Term ( ) 0.90 Amt Choudhary et al/(ijcse) Internatonal Journal on Computer Scence and Engneerng, Maxmum Epochs 000 Termnaton Condtons Based on mnmum Mean Square Error or maxmum number of epochs allowed Intal Weghts and based term values Randomly generated values between 0 and Number of Hdden Layers V. FUNCTIONAL DETAILS The neural networ structure s determned by the number of neurons n the nput, hdden and output layers. The networ performance s also nfluenced by the actvaton functons of the hdden and output layer neurons. The number of neurons n the nput and output layer of the networ depends on the type of the problem. In the current stuaton, the number of neurons n the nput and output layers are fxed at 00 and 0 respectvely. The number of neurons n the hdden layer and the actvaton functons of the neurons n the hdden and output layers are to be decded. It s very dffcult to determne the optmal number of hdden neurons. Too few hdden neurons wll result n underfttng and there wll be hgh tranng error and statstcal error due to the lac of enough adjustable parameters to map the nput-output relatonshp. Too many hdden neurons wll result n overfttng and hgh varance. The networ wll tend to memorse the nput-output relatons and normally fal to generalze. Testng data or unseen data could not be mapped properly. Each neuron n the neural networ has a transformaton functon. To produce an output, the neuron performs the transformaton functon on the weghted sum of ts nputs. Varous transfer functons used n neural networs are compet, hardlm, logsg, posln, pureln, radbas, satln, softmax, tansg and trbas etc. The three transfer functons normally used n MLP are logsg, tansg and pureln. logsg transfer functon s also nown as Logstc Sgmod: log sg( x), x e tansg transfer functon s also nown as hyperbolc tangent: x x e e tan sg( x) x x e e The hyperbolc tangent and logstc sgmod are related by: tan sg( x) 2 e We get: 2x 2 tan sg ( x) 2x e pureln transfer functon s also nown as smple lnear transfer functon and produces the same output as ts nput. pureln ( x) x 4
5 These transfer functons are commonly used as they are easy to use mathematcally and whle saturatng, they are close to lnear near the orgn. There s not any rule that gves the calculaton for the deal parameter settng for a neural networ. In our problem, the structure analyss has to be carred out to fnd the optmum number of neurons n the hdden layer and the best actvaton functons for the neurons n the hdden and output layers. For ths analyss, the actvaton functon of the neurons n the hdden layer was selected as logsg and the actvaton functon of the neurons n the output layer was changed and the process s repeated by selectng dfferent number of neurons n the hdden layer as shown n Table. Smlarly, as shown n Table 2 and Table 3, the actvaton functon of the hdden layer neurons was selected as tansg and pureln respectvely and the actvaton functon s changed for output layer neurons wth varous dfferent numbers of neurons n the hdden layer. Mean Square Error (MSE) s an accepted measure of the performance ndex often used n MLP networs. The lower value of MSE ndcates that the networ s capable of mappng the nput and output accurately. The accepted error level s set to 0.00 and the tranng wll stop when the fnal value of MSE reaches at 0.00 or below ths level. The number of epochs requred to tran a networ also ndcates the networ performance. The adjustable parameters of the networ wll not converge properly f the number of tranng epochs s nsuffcent and the networ wll not be well traned. On the other hand, the networ wll tae unnecessary long tranng tme f there are excessve tranng epochs. The number of tranng epochs should be suffcent enough so as to meet the am of tranng. If there s a saturated or horzontal straght lne at the end of the MSE plot, the further tranng epochs are no longer benefcal and the tranng should be stopped by ntroducng the stoppng crteron such as maxmum number of tranng epochs allowed n addton to the stoppng crteron of maxmum acceptable error as specfed n the tranng algorthm. Ths type of MSE plot s observed when there s a very complcated problem or nsuffcent tranng algorthm or the networ have very lmted resources. Amt Choudhary et al/(ijcse) Internatonal Journal on Computer Scence and Engneerng, networ decreases. Insuffcent number of hdden neurons results n underfttng and the MSE does not fall below the acceptable error level and the tranng s stopped by the stoppng crteron of maxmum allowed number of tranng epochs. Overfttng s observed when the number of hdden neurons exceeds a certan count and the percentage recognton accuracy gradually falls wth the ncrease n the number of hdden neurons. The networ wll tend to memorse the nputoutput relatons and normally fal to generalze. However the tranng of the networ becomes faster as ndcated by the decrease n number of tranng epochs as the number of hdden neurons ncreases. The optmal number of hdden neurons and the best combnaton of actvaton functons for the hdden and output neurons can be decded by consderng the numeral recognton accuracy of the networ. By detaled analyss of Table3, Table 4 and Table 5, t s clear that the good recognton accuracy s observed when the selected combnaton of actvaton functon for hdden and output neurons were logsg-tansg, logsgpurelne, tansg-tansg and tansg-purelne. By tang nto consderaton all the performance measure parameters such as MSE, overfttng and underfttng, tranng tme and the avalable system resources, t s evdent from Table 2 that the neural networ wth 25 neurons n the hdden layer and wth tansg-tansg actvaton functon combnaton for the hdden and output neurons provdes the best sample recognton accuracy of 99.% wth least MSE of and suffcent number of tranng epochs. For ths networ, the profle of MSE plot for the tranng epochs s drawn n Fg 5 VI. RESULTS AND DISCUSSION Three types of transfer functons: logsg, tansg and pureln are used to smulate the neural networ used n our problem of numeral recognton. As seen n Table 3, the actvaton functon for the output layer was changed to logsg, tansg and pureln one by one by fxng the actvaton functon of hdden neurons to logsg. The number of tranng epochs and MSE values of the networ are ndcated correspondng to the number of neurons n the hdden layer. Smlarly n Table 4 and Table 5, the actvaton functon of the hdden neurons was fxed as tansg and pureln respectvely. If we gnore the type of actvaton functon used n hdden layer and output layer, t s clear from Table 3, Table 4 and Table 5, that, as the number of neurons n the hdden layer s ncreased, the number of epochs requred for the tranng of the Fgure 5. The varaton of MSE wth the tranng epochs VII. CONCLUSION AND FUTURE SCOPE The MLP used n the proposed method for the handwrtten numeral recognton employng the bac propagaton algorthm performed exceptonally well wth 25 neurons n the hdden layer and tansg as the actvaton functon for both hdden and output layer neurons. In ths optmal MLP structure, the number of neurons n the nput and output layers were already fxed to 00 and 0 respectvely. 5
6 Thus t can be concluded that f the accuracy of the results s a crtcal factor for an applcaton, then the networ havng the optmal structure should be used but f tranng tme s a crtcal factor then the networ havng a large number of hdden neurons should be used. Tranng speed can be acheved at the cost of system resources because the number of tranng epochs decreases wth the ncrease n the number of hdden neurons. More wor need to be done especally on the test for more complex handwrtten numerals. The proposed wor can be carred out to recognze numeral strngs after proper segmentaton of the dgts nto solated dgt mages. REFERENCES [] S Srhar, V Govndraju, R Srhar. Handwrtten text recognton. Proc 4 th Internatonal Worshop on fronters n handwrtng recognton, Tawan, , 994. [2] Neves, R.F.P., et al A New Technque to Threshold the Courtesy Amount of Brazlan Ban Checs. In Proceedngs of the 5th IWSSIP, (Bratslava, Slova Republc, June 2008), IEEE Press, n press. [3] S. Ouchtat, B. Mould, A. Lachour, Segmentaton and Recognton of Handwrtten Numerc Chans. In Journal of Computer Scence 3 (4) ISSN , , 997. Amt Choudhary et al/(ijcse) Internatonal Journal on Computer Scence and Engneerng, [4] N. Arca and F. Yarman-Vural, "An Overvew of Character Recognton Focused on Offlne Handwrtng", IEEE Transactons on Systems, Man, and Cybernetcs, Part C: Applcatons and Revews, Vol. 3, No.2, pp , 200. [5] B. K. Verma, "Handwrtten Hnd Character Recognton Usng Multlayer Perceptron and Radal Bass Functon Neural Networ", IEEE Internatonal Conference on Neural Networ, vol. 4, pp , 995. [6] R. Iln; R. Kozma; P. J. Werbos. Beyond feedforward models traned by bacpropagaton: A practcal tranng tool for a more effcent unversal approxmator. IEEE Transactons on Neural Networs, 9(6): , June [7] An OCR System for Telugu, Proceedngs of the Sxth Internatonal Conference on Document Analyss and Recognton, p.0, September 0-3, 200 [8] N. P. Banashree, R. Vasanta, OCR for Scrpt Identfcaton of Hnd (Devnagar) Numerals usng Feature Sub Selecton by Means of End- Pont wth Neuro-Memetc Model. In Proceedngs of World Academy of Scence, Engneerng and Technology, vol. 22 ISSN , 78 82, [9] B. Burton, R.G. Harley, G. Dana, and J.R. Rodgerson, "Reducng the Computatonal Demands of Contnually Onlne- Traned Artfcal Neural Networs for System Identfcaton and Control of Fast Processes" IEEE transacton on Industry Applcatons,Vol 34. no.3, May/June 998, pp TABLE III. BEHAVIOR OF MLP WITH LOGSIG FUNCTION IN HIDDEN LAYER Number of Logsg- Logsg Logsg -Tansg Logsg -Purelne Hdden Unts % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % 6
7 TABLE IV. Amt Choudhary et al/(ijcse) Internatonal Journal on Computer Scence and Engneerng, BEHAVIOR OF MLP WITH TANSIG FUNCTION IN HIDDEN LAYER Number of Tansg - Logsg Tansg Tansg Tansg -Purelne Hdden Unts % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % TABLE V. BEHAVIOR OF MLP WITH PURELIN FUNCTION IN HIDDEN LAYER Number of Purelne - Logsg Purelne Tansg Purelne Purelne Hdden Unts % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % 7
Lecture 2: Single Layer Perceptrons Kevin Swingler
Lecture 2: Sngle Layer Perceptrons Kevn Sngler kms@cs.str.ac.uk Recap: McCulloch-Ptts Neuron Ths vastly smplfed model of real neurons s also knon as a Threshold Logc Unt: W 2 A Y 3 n W n. A set of synapses
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 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 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 informationStatistical Approach for Offline Handwritten Signature Verification
Journal of Computer Scence 4 (3): 181-185, 2008 ISSN 1549-3636 2008 Scence Publcatons Statstcal Approach for Offlne Handwrtten Sgnature Verfcaton 2 Debnath Bhattacharyya, 1 Samr Kumar Bandyopadhyay, 2
More informationSPEE Recommended Evaluation Practice #6 Definition of Decline Curve Parameters Background:
SPEE Recommended Evaluaton Practce #6 efnton of eclne Curve Parameters Background: The producton hstores of ol and gas wells can be analyzed to estmate reserves and future ol and gas producton rates and
More information"Research Note" APPLICATION OF CHARGE SIMULATION METHOD TO ELECTRIC FIELD CALCULATION IN THE POWER CABLES *
Iranan Journal of Scence & Technology, Transacton B, Engneerng, ol. 30, No. B6, 789-794 rnted n The Islamc Republc of Iran, 006 Shraz Unversty "Research Note" ALICATION OF CHARGE SIMULATION METHOD TO ELECTRIC
More informationLogistic Regression. Lecture 4: More classifiers and classes. Logistic regression. Adaboost. Optimization. Multiple class classification
Lecture 4: More classfers and classes C4B Machne Learnng Hlary 20 A. Zsserman Logstc regresson Loss functons revsted Adaboost Loss functons revsted Optmzaton Multple class classfcaton Logstc Regresson
More informationA hybrid global optimization algorithm based on parallel chaos optimization and outlook algorithm
Avalable onlne www.ocpr.com Journal of Chemcal and Pharmaceutcal Research, 2014, 6(7):1884-1889 Research Artcle ISSN : 0975-7384 CODEN(USA) : JCPRC5 A hybrd global optmzaton algorthm based on parallel
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 informationBiometric Signature Processing & Recognition Using Radial Basis Function Network
Bometrc Sgnature Processng & Recognton Usng Radal Bass Functon Network Ankt Chadha, Neha Satam, and Vbha Wal Abstract- Automatc recognton of sgnature s a challengng problem whch has receved much attenton
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 information8.5 UNITARY AND HERMITIAN MATRICES. The conjugate transpose of a complex matrix A, denoted by A*, is given by
6 CHAPTER 8 COMPLEX VECTOR SPACES 5. Fnd the kernel of the lnear transformaton gven n Exercse 5. In Exercses 55 and 56, fnd the mage of v, for the ndcated composton, where and are gven by the followng
More informationFace Verification Problem. Face Recognition Problem. Application: Access Control. Biometric Authentication. Face Verification (1:1 matching)
Face Recognton Problem Face Verfcaton Problem Face Verfcaton (1:1 matchng) Querymage face query Face Recognton (1:N matchng) database Applcaton: Access Control www.vsage.com www.vsoncs.com Bometrc Authentcaton
More informationAn Enhanced Super-Resolution System with Improved Image Registration, Automatic Image Selection, and Image Enhancement
An Enhanced Super-Resoluton System wth Improved Image Regstraton, Automatc Image Selecton, and Image Enhancement Yu-Chuan Kuo ( ), Chen-Yu Chen ( ), and Chou-Shann Fuh ( ) Department of Computer Scence
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 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 informationv a 1 b 1 i, a 2 b 2 i,..., a n b n i.
SECTION 8.4 COMPLEX VECTOR SPACES AND INNER PRODUCTS 455 8.4 COMPLEX VECTOR SPACES AND INNER PRODUCTS All the vector spaces we have studed thus far n the text are real vector spaces snce the scalars are
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 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 informationGender Classification for Real-Time Audience Analysis System
Gender Classfcaton for Real-Tme Audence Analyss System Vladmr Khryashchev, Lev Shmaglt, Andrey Shemyakov, Anton Lebedev Yaroslavl State Unversty Yaroslavl, Russa vhr@yandex.ru, shmaglt_lev@yahoo.com, andrey.shemakov@gmal.com,
More informationA Set of Features Extraction Methods for the Recognition of the Isolated Handwritten Digits
Internatonal Journal of Computer and Communcaton Engneerng, Vol. 3, No. 5, September 214 A Set of Features Extracton Methods for the Recognton of the Isolated Handwrtten Dgts S. Ouchtat, M. Redjm, and
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 informationSingle and multiple stage classifiers implementing logistic discrimination
Sngle and multple stage classfers mplementng logstc dscrmnaton Hélo Radke Bttencourt 1 Dens Alter de Olvera Moraes 2 Vctor Haertel 2 1 Pontfíca Unversdade Católca do Ro Grande do Sul - PUCRS Av. Ipranga,
More informationL10: Linear discriminants analysis
L0: Lnear dscrmnants analyss Lnear dscrmnant analyss, two classes Lnear dscrmnant analyss, C classes LDA vs. PCA Lmtatons of LDA Varants of LDA Other dmensonalty reducton methods CSCE 666 Pattern Analyss
More informationA machine vision approach for detecting and inspecting circular parts
A machne vson approach for detectng and nspectng crcular parts Du-Mng Tsa Machne Vson Lab. Department of Industral Engneerng and Management Yuan-Ze Unversty, Chung-L, Tawan, R.O.C. E-mal: edmtsa@saturn.yzu.edu.tw
More informationOn the Optimal Control of a Cascade of Hydro-Electric Power Stations
On the Optmal Control of a Cascade of Hydro-Electrc Power Statons M.C.M. Guedes a, A.F. Rbero a, G.V. Smrnov b and S. Vlela c a Department of Mathematcs, School of Scences, Unversty of Porto, Portugal;
More informationRESEARCH ON DUAL-SHAKER SINE VIBRATION CONTROL. Yaoqi FENG 1, Hanping QIU 1. China Academy of Space Technology (CAST) yaoqi.feng@yahoo.
ICSV4 Carns Australa 9- July, 007 RESEARCH ON DUAL-SHAKER SINE VIBRATION CONTROL Yaoq FENG, Hanpng QIU Dynamc Test Laboratory, BISEE Chna Academy of Space Technology (CAST) yaoq.feng@yahoo.com Abstract
More informationProject Networks With Mixed-Time Constraints
Project Networs Wth Mxed-Tme Constrants L Caccetta and B Wattananon Western Australan Centre of Excellence n Industral Optmsaton (WACEIO) Curtn Unversty of Technology GPO Box U1987 Perth Western Australa
More informationECE544NA Final Project: Robust Machine Learning Hardware via Classifier Ensemble
1 ECE544NA Fnal Project: Robust Machne Learnng Hardware va Classfer Ensemble Sa Zhang, szhang12@llnos.edu Dept. of Electr. & Comput. Eng., Unv. of Illnos at Urbana-Champagn, Urbana, IL, USA Abstract In
More informationMATHEMATICAL ENGINEERING TECHNICAL REPORTS. Sequential Optimizing Investing Strategy with Neural Networks
MATHEMATICAL ENGINEERING TECHNICAL REPORTS Sequental Optmzng Investng Strategy wth Neural Networks Ryo ADACHI and Akmch TAKEMURA METR 2010 03 February 2010 DEPARTMENT OF MATHEMATICAL INFORMATICS GRADUATE
More informationAn interactive system for structure-based ASCII art creation
An nteractve system for structure-based ASCII art creaton Katsunor Myake Henry Johan Tomoyuk Nshta The Unversty of Tokyo Nanyang Technologcal Unversty Abstract Non-Photorealstc Renderng (NPR), whose am
More informationOffline Verification of Hand Written Signature using Adaptive Resonance Theory Net (Type-1)
Internatonal Journal of Sgnal Processng Systems Vol, No June 203 Offlne Verfcaton of Hand Wrtten Sgnature usng Adaptve Resonance Theory Net (Type-) Trtharaj Dash Veer Surendra Sa Unversty of Technology,
More informationSupport Vector Machines
Support Vector Machnes Max Wellng Department of Computer Scence Unversty of Toronto 10 Kng s College Road Toronto, M5S 3G5 Canada wellng@cs.toronto.edu Abstract Ths s a note to explan support vector machnes.
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 informationAn artificial Neural Network approach to monitor and diagnose multi-attribute quality control processes. S. T. A. Niaki*
Journal of Industral Engneerng Internatonal July 008, Vol. 4, No. 7, 04 Islamc Azad Unversty, South Tehran Branch An artfcal Neural Network approach to montor and dagnose multattrbute qualty control processes
More informationCausal, Explanatory Forecasting. Analysis. Regression Analysis. Simple Linear Regression. Which is Independent? Forecasting
Causal, Explanatory Forecastng Assumes cause-and-effect relatonshp between system nputs and ts output Forecastng wth Regresson Analyss Rchard S. Barr Inputs System Cause + Effect Relatonshp The job of
More informationCHAPTER 5 RELATIONSHIPS BETWEEN QUANTITATIVE VARIABLES
CHAPTER 5 RELATIONSHIPS BETWEEN QUANTITATIVE VARIABLES In ths chapter, we wll learn how to descrbe the relatonshp between two quanttatve varables. Remember (from Chapter 2) that the terms quanttatve varable
More informationOptimal Choice of Random Variables in D-ITG Traffic Generating Tool using Evolutionary Algorithms
Optmal Choce of Random Varables n D-ITG Traffc Generatng Tool usng Evolutonary Algorthms M. R. Mosav* (C.A.), F. Farab* and S. Karam* Abstract: Impressve development of computer networks has been requred
More informationConversion between the vector and raster data structures using Fuzzy Geographical Entities
Converson between the vector and raster data structures usng Fuzzy Geographcal Enttes Cdála Fonte Department of Mathematcs Faculty of Scences and Technology Unversty of Combra, Apartado 38, 3 454 Combra,
More informationPerformance Analysis and Coding Strategy of ECOC SVMs
Internatonal Journal of Grd and Dstrbuted Computng Vol.7, No. (04), pp.67-76 http://dx.do.org/0.457/jgdc.04.7..07 Performance Analyss and Codng Strategy of ECOC SVMs Zhgang Yan, and Yuanxuan Yang, School
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 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 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 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 informationPortfolio Loss Distribution
Portfolo Loss Dstrbuton Rsky assets n loan ortfolo hghly llqud assets hold-to-maturty n the bank s balance sheet Outstandngs The orton of the bank asset that has already been extended to borrowers. Commtment
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 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 informationVision Mouse. Saurabh Sarkar a* University of Cincinnati, Cincinnati, USA ABSTRACT 1. INTRODUCTION
Vson Mouse Saurabh Sarkar a* a Unversty of Cncnnat, Cncnnat, USA ABSTRACT The report dscusses a vson based approach towards trackng of eyes and fngers. The report descrbes the process of locatng the possble
More informationOn the Use of Neural Network as a Universal Approximator
Internatonal Journal of Scences and Technques of Automatc control & computer engneerng IJ-STA, Volume, N, Jul 8, pp 386 399 On the Use of Neural Network as a Unversal Appromator Amel SIFAOUI, Afef ABDELKRIM,
More information1. Fundamentals of probability theory 2. Emergence of communication traffic 3. Stochastic & Markovian Processes (SP & MP)
6.3 / -- Communcaton Networks II (Görg) SS20 -- www.comnets.un-bremen.de Communcaton Networks II Contents. Fundamentals of probablty theory 2. Emergence of communcaton traffc 3. Stochastc & Markovan Processes
More informationDamage detection in composite laminates using coin-tap method
Damage detecton n composte lamnates usng con-tap method S.J. Km Korea Aerospace Research Insttute, 45 Eoeun-Dong, Youseong-Gu, 35-333 Daejeon, Republc of Korea yaeln@kar.re.kr 45 The con-tap test has the
More informationChapter 4 ECONOMIC DISPATCH AND UNIT COMMITMENT
Chapter 4 ECOOMIC DISATCH AD UIT COMMITMET ITRODUCTIO A power system has several power plants. Each power plant has several generatng unts. At any pont of tme, the total load n the system s met by the
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 study on the ability of Support Vector Regression and Neural Networks to Forecast Basic Time Series Patterns
A study on the ablty of Support Vector Regresson and Neural Networks to Forecast Basc Tme Seres Patterns Sven F. Crone, Jose Guajardo 2, and Rchard Weber 2 Lancaster Unversty, Department of Management
More informationFault tolerance in cloud technologies presented as a service
Internatonal Scentfc Conference Computer Scence 2015 Pavel Dzhunev, PhD student Fault tolerance n cloud technologes presented as a servce INTRODUCTION Improvements n technques for vrtualzaton and performance
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 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 informationAutomated Network Performance Management and Monitoring via One-class Support Vector Machine
Automated Network Performance Management and Montorng va One-class Support Vector Machne R. Zhang, J. Jang, and S. Zhang Dgtal Meda & Systems Research Insttute, Unversty of Bradford, UK Abstract: In ths
More informationModelling of Web Domain Visits by Radial Basis Function Neural Networks and Support Vector Machine Regression
Modellng of Web Doman Vsts by Radal Bass Functon Neural Networks and Support Vector Machne Regresson Vladmír Olej, Jana Flpová Insttute of System Engneerng and Informatcs Faculty of Economcs and Admnstraton,
More informationInvestigation of Normalization Techniques and Their Impact on a Recognition Rate in Handwritten Numeral Recognition
S C H E D A E I N F O R M A T I C A E VOLUME 19 010 Investgaton of Normalzaton Technques and Ther Impact on a Recognton Rate n Handwrtten Numeral Recognton WIESŁAW CHMIELNICKI 1, KATARZYNA STĄPOR 1 Faculty
More informationDevelopment of an intelligent system for tool wear monitoring applying neural networks
of Achevements n Materals and Manufacturng Engneerng VOLUME 14 ISSUE 1-2 January-February 2006 Development of an ntellgent system for tool wear montorng applyng neural networks A. Antć a, J. Hodolč a,
More informationA Multi-mode Image Tracking System Based on Distributed Fusion
A Mult-mode Image Tracng System Based on Dstrbuted Fuson Ln zheng Chongzhao Han Dongguang Zuo Hongsen Yan School of Electroncs & nformaton engneerng, X an Jaotong Unversty X an, Shaanx, Chna Lnzheng@malst.xjtu.edu.cn
More informationA Simple Approach to Clustering in Excel
A Smple Approach to Clusterng n Excel Aravnd H Center for Computatonal Engneerng and Networng Amrta Vshwa Vdyapeetham, Combatore, Inda C Rajgopal Center for Computatonal Engneerng and Networng Amrta Vshwa
More informationDesign and Development of a Security Evaluation Platform Based on International Standards
Internatonal Journal of Informatcs Socety, VOL.5, NO.2 (203) 7-80 7 Desgn and Development of a Securty Evaluaton Platform Based on Internatonal Standards Yuj Takahash and Yoshm Teshgawara Graduate School
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 informationForensic Handwritten Document Retrieval System
Forensc Handwrtten Document Retreval System Sargur N SRIHARI and Zhxn SHI + Center of Excellence for Document Analyss and Recognton (CEDAR), Unversty at Buffalo, State Unversty of New York, Buffalo, USA
More informationAn MILP model for planning of batch plants operating in a campaign-mode
An MILP model for plannng of batch plants operatng n a campagn-mode Yanna Fumero Insttuto de Desarrollo y Dseño CONICET UTN yfumero@santafe-concet.gov.ar Gabrela Corsano Insttuto de Desarrollo y Dseño
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 informationDesign of Output Codes for Fast Covering Learning using Basic Decomposition Techniques
Journal of Computer Scence (7): 565-57, 6 ISSN 59-66 6 Scence Publcatons Desgn of Output Codes for Fast Coverng Learnng usng Basc Decomposton Technques Aruna Twar and Narendra S. Chaudhar, Faculty of Computer
More informationAutomated information technology for ionosphere monitoring of low-orbit navigation satellite signals
Automated nformaton technology for onosphere montorng of low-orbt navgaton satellte sgnals Alexander Romanov, Sergey Trusov and Alexey Romanov Federal State Untary Enterprse Russan Insttute of Space Devce
More informationTesting CAB-IDS through Mutations: on the Identification of Network Scans
Testng CAB-IDS through Mutatons: on the Identfcaton of Network Scans Emlo Corchado, Álvaro Herrero, José Manuel Sáz Department of Cvl Engneerng, Unversty of Burgos, Span {escorchado, ahcoso, msaz}@ubu.es
More informationSOLVING CARDINALITY CONSTRAINED PORTFOLIO OPTIMIZATION PROBLEM BY BINARY PARTICLE SWARM OPTIMIZATION ALGORITHM
SOLVIG CARDIALITY COSTRAIED PORTFOLIO OPTIMIZATIO PROBLEM BY BIARY PARTICLE SWARM OPTIMIZATIO ALGORITHM Aleš Kresta Klíčová slova: optmalzace portfola, bnární algortmus rojení částc Key words: portfolo
More informationData Visualization by Pairwise Distortion Minimization
Communcatons n Statstcs, Theory and Methods 34 (6), 005 Data Vsualzaton by Parwse Dstorton Mnmzaton By Marc Sobel, and Longn Jan Lateck* Department of Statstcs and Department of Computer and Informaton
More informationPOLYSA: A Polynomial Algorithm for Non-binary Constraint Satisfaction Problems with and
POLYSA: A Polynomal Algorthm for Non-bnary Constrant Satsfacton Problems wth and Mguel A. Saldo, Federco Barber Dpto. Sstemas Informátcos y Computacón Unversdad Poltécnca de Valenca, Camno de Vera s/n
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 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 informationHow Sets of Coherent Probabilities May Serve as Models for Degrees of Incoherence
1 st Internatonal Symposum on Imprecse Probabltes and Ther Applcatons, Ghent, Belgum, 29 June 2 July 1999 How Sets of Coherent Probabltes May Serve as Models for Degrees of Incoherence Mar J. Schervsh
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 informationBERNSTEIN POLYNOMIALS
On-Lne Geometrc Modelng Notes BERNSTEIN POLYNOMIALS Kenneth I. Joy Vsualzaton and Graphcs Research Group Department of Computer Scence Unversty of Calforna, Davs Overvew Polynomals are ncredbly useful
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 informationLogical Development Of Vogel s Approximation Method (LD-VAM): An Approach To Find Basic Feasible Solution Of Transportation Problem
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME, ISSUE, FEBRUARY ISSN 77-866 Logcal Development Of Vogel s Approxmaton Method (LD- An Approach To Fnd Basc Feasble Soluton Of Transportaton
More informationINVESTIGATION OF VEHICULAR USERS FAIRNESS IN CDMA-HDR NETWORKS
21 22 September 2007, BULGARIA 119 Proceedngs of the Internatonal Conference on Informaton Technologes (InfoTech-2007) 21 st 22 nd September 2007, Bulgara vol. 2 INVESTIGATION OF VEHICULAR USERS FAIRNESS
More informationNPAR TESTS. One-Sample Chi-Square Test. Cell Specification. Observed Frequencies 1O i 6. Expected Frequencies 1EXP i 6
PAR TESTS If a WEIGHT varable s specfed, t s used to replcate a case as many tmes as ndcated by the weght value rounded to the nearest nteger. If the workspace requrements are exceeded and samplng has
More informationFast Fuzzy Clustering of Web Page Collections
Fast Fuzzy Clusterng of Web Page Collectons Chrstan Borgelt and Andreas Nürnberger Dept. of Knowledge Processng and Language Engneerng Otto-von-Guercke-Unversty of Magdeburg Unverstätsplatz, D-396 Magdeburg,
More informationFrequency Selective IQ Phase and IQ Amplitude Imbalance Adjustments for OFDM Direct Conversion Transmitters
Frequency Selectve IQ Phase and IQ Ampltude Imbalance Adjustments for OFDM Drect Converson ransmtters Edmund Coersmeer, Ernst Zelnsk Noka, Meesmannstrasse 103, 44807 Bochum, Germany edmund.coersmeer@noka.com,
More informationAdaptive Fractal Image Coding in the Frequency Domain
PROCEEDINGS OF INTERNATIONAL WORKSHOP ON IMAGE PROCESSING: THEORY, METHODOLOGY, SYSTEMS AND APPLICATIONS 2-22 JUNE,1994 BUDAPEST,HUNGARY Adaptve Fractal Image Codng n the Frequency Doman K AI UWE BARTHEL
More informationMooring Pattern Optimization using Genetic Algorithms
6th World Congresses of Structural and Multdscplnary Optmzaton Ro de Janero, 30 May - 03 June 005, Brazl Moorng Pattern Optmzaton usng Genetc Algorthms Alonso J. Juvnao Carbono, Ivan F. M. Menezes Luz
More informationHybrid-Learning Methods for Stock Index Modeling
Hybrd-Learnng Methods for Stock Index Modelng 63 Chapter IV Hybrd-Learnng Methods for Stock Index Modelng Yuehu Chen, Jnan Unversty, Chna Ajth Abraham, Chung-Ang Unversty, Republc of Korea Abstract The
More informationDropout: A Simple Way to Prevent Neural Networks from Overfitting
Journal of Machne Learnng Research 15 (2014) 1929-1958 Submtted 11/13; Publshed 6/14 Dropout: A Smple Way to Prevent Neural Networks from Overfttng Ntsh Srvastava Geoffrey Hnton Alex Krzhevsky Ilya Sutskever
More informationA Binary Particle Swarm Optimization Algorithm for Lot Sizing Problem
Journal o Economc and Socal Research 5 (2), -2 A Bnary Partcle Swarm Optmzaton Algorthm or Lot Szng Problem M. Fath Taşgetren & Yun-Cha Lang Abstract. Ths paper presents a bnary partcle swarm optmzaton
More informationLoop Parallelization
- - Loop Parallelzaton C-52 Complaton steps: nested loops operatng on arrays, sequentell executon of teraton space DECLARE B[..,..+] FOR I :=.. FOR J :=.. I B[I,J] := B[I-,J]+B[I-,J-] ED FOR ED FOR analyze
More informationRing structure of splines on triangulations
www.oeaw.ac.at Rng structure of splnes on trangulatons N. Vllamzar RICAM-Report 2014-48 www.rcam.oeaw.ac.at RING STRUCTURE OF SPLINES ON TRIANGULATIONS NELLY VILLAMIZAR Introducton For a trangulated regon
More informationResearch Article Enhanced Two-Step Method via Relaxed Order of α-satisfactory Degrees for Fuzzy Multiobjective Optimization
Hndaw Publshng Corporaton Mathematcal Problems n Engneerng Artcle ID 867836 pages http://dxdoorg/055/204/867836 Research Artcle Enhanced Two-Step Method va Relaxed Order of α-satsfactory Degrees for Fuzzy
More informationBinomial Link Functions. Lori Murray, Phil Munz
Bnomal Lnk Functons Lor Murray, Phl Munz Bnomal Lnk Functons Logt Lnk functon: ( p) p ln 1 p Probt Lnk functon: ( p) 1 ( p) Complentary Log Log functon: ( p) ln( ln(1 p)) Motvatng Example A researcher
More informationOpen Access A Load Balancing Strategy with Bandwidth Constraint in Cloud Computing. Jing Deng 1,*, Ping Guo 2, Qi Li 3, Haizhu Chen 1
Send Orders for Reprnts to reprnts@benthamscence.ae The Open Cybernetcs & Systemcs Journal, 2014, 8, 115-121 115 Open Access A Load Balancng Strategy wth Bandwdth Constrant n Cloud Computng Jng Deng 1,*,
More informationDescriptive Models. Cluster Analysis. Example. General Applications of Clustering. Examples of Clustering Applications
CMSC828G Prncples of Data Mnng Lecture #9 Today s Readng: HMS, chapter 9 Today s Lecture: Descrptve Modelng Clusterng Algorthms Descrptve Models model presents the man features of the data, a global summary
More informationThe Effect of Mean Stress on Damage Predictions for Spectral Loading of Fiberglass Composite Coupons 1
EWEA, Specal Topc Conference 24: The Scence of Makng Torque from the Wnd, Delft, Aprl 9-2, 24, pp. 546-555. The Effect of Mean Stress on Damage Predctons for Spectral Loadng of Fberglass Composte Coupons
More informationTime Domain simulation of PD Propagation in XLPE Cables Considering Frequency Dependent Parameters
Internatonal Journal of Smart Grd and Clean Energy Tme Doman smulaton of PD Propagaton n XLPE Cables Consderng Frequency Dependent Parameters We Zhang a, Jan He b, Ln Tan b, Xuejun Lv b, Hong-Je L a *
More informationHow To Know The Components Of Mean Squared Error Of Herarchcal Estmator S
S C H E D A E I N F O R M A T I C A E VOLUME 0 0 On Mean Squared Error of Herarchcal Estmator Stans law Brodowsk Faculty of Physcs, Astronomy, and Appled Computer Scence, Jagellonan Unversty, Reymonta
More information