Handling Uncertain Spatial Data: Comparisons between Indexing Structures. Bir Bhanu, Rui Li, Chinya Ravishankar and Jinfeng Ni
|
|
- Adam Hodges
- 6 years ago
- Views:
Transcription
1 Handlng Uncertan Spatal Data: Comparsons between Indexng Structures Br Bhanu, Ru L, Chnya Ravshankar and Jnfeng N Abstract Managng and manpulatng uncertanty n spatal databases are mportant problems for varous practcal applcatons. Unlke the tradtonal fuzzy approaches n relatonal databases, n ths paper we propose a probablty-based method to model and ndex uncertan spatal data where every object s represented by a probablty densty functon (PDF). To ndex PDFs, we construct an optmzed Gaussan mxture herarchy () and two varants of uncertan R-tree. We provde a comprehensve comparson among these three ndces and plan R-tree on TIGER/Lne Southern Calforna landmark pont dataset. We fnd that uncertan R-tree s the best for fxed query and s sutable for both certan and uncertan queres. Moreover, s sutable not only for spatal databases, but also for mult-dmensonal ndexng applcatons lke content based mage retreval, where R-tree s neffcent n hgh dmensons. Index Terms ndex structures, R-tree, spatal database, uncertanty. I. INTRODUCTION Geographc nformaton system (GIS) s a system of computer software, hardware and data, and personnel to help manpulate, analyze and present nformaton that s ted to a spatal locaton. Spatal database s the system whch organzes spatal nformaton n GIS []. In GIS applcatons, t s generally agreed that there are several types of error (uncertanty) whch determne the overall accuracy of fnal products. A bennal Spatal Accuracy Symposum s held specfcally on ths topc []. However, n the fled of nformaton technology scant attenton has been pad for handlng uncertanty n spatal databases [3]. Even though there s more awareness and some understandng of uncertanty sources n spatal data, at present there s not a complete system whch can store and operate on uncertan Ths work was supported by NSF Informaton Technology Research Grant The contents of the nformaton do not reflect the poston or polcy of the U.S. Government. B. Bhanu s wth the Center for Research n Intellgent System, Unversty of Calforna, Rversde, CA 5 (e-mal: bhanu@crs.ucr.edu). R. L s wth the Department of Electrcal Engneerng, Unversty of Calforna, Rversde, CA 95 (e-mal: rl@vslab.ucr.edu). C. Ravshankar s wth the Department of Computer Scence & Engneerng, Unversty of Calforna, Rversde, CA 95 (e-mal: rav@cs.ucr.edu). J. N s wth the Department of Computer Scence & Engneerng, Unversty of Calforna, Rversde, CA 95 (e-mal: jn@cs.ucr.edu). spatal data. ArcInfo and Oracle Extensons are only suted for certan spatal data []. In ths paper, we present our approaches on ndexng uncertan spatal data. Most of the exstng approaches for management of probablstc data are based on the relatonal model and use fuzzy set theory [4, 5]. They are useful for representng uncertanty at the symbolc level. However, n addton to symbolc uncertanty, sensor-processng tasks nvolve uncertantes at both numerc and exstence levels. Supportng these types of uncertanty n the current relatonal model usng fuzzy logc s fundamentally dffcult. So n our approach, we use probablty densty functons (PDFs) to represent uncertan data, whch means every object s a random varable. In spatal databases, R-tree s the most often used ndexng structure, whch s a depth-balanced tree whose nodes are represented by Mnmum Boundng Rectangles (MBR). Guttman [6] frst proposed R-tree and Greene and Beckmann [7, 8] optmzed t by reducng margn and MBR overlap. But R-tree famly ndexes fxed data only. In order to handle the uncertan nformaton related to each object, we have to use R- tree varant or desgn new ndex. In ths paper, we buld two varants of uncertan R-tree and an optmzed Gaussan mxture herarchy () based on Gaussan mxture model (GMM). The rest of the paper s organzed as follows. The uncertanty representaton and ndex constructon are explaned n Secton II. Secton III gves the expermental results on TIGER/Lne Southern Calforna landmark pont dataset. Secton IV concludes the paper. A. PDF representaton II. TECHNICAL APPROACH In conventonal spatal databases, each object s represented by a feature vector n n dmensonal feature space. But when the data are uncertan, a dfferent representaton s requred. In out approach, we use a PDF to represent each uncertan object. If an object needs n features to descrbe, then t s an n dmensonal (feature vector) random varable, as gven n (). where f =,, N [ f f,, f ] T, =, N s the number of objects For smplcty, we assume that the features are ndependent of each other and each feature s PDF ( f j, j =,..., n ) s known. The process of gettng the PDFs s called uncertanty n ()
2 modelng and t s a part of our on gong related work. Our system needs to handle both certan and uncertan data. So we do not gve a specfc name to uncertan data. Feature vector s the name for both certan and uncertan data n ths paper. For feature vectors, metrcs lke Eucldean dstance, Manhattan dstance etc. are used to measure smlarty. Snce our uncertan objects are random varables represented by PDFs, we defne the smlarty of feature vectors as the probablty that two random varables are the same, as shown n (). In ths equaton, D and Q are two objects. D s the object n the database and Q s the query. s a threshold vector descrbng the maxmal error the system can tolerate to stll regard D as smlar to Q. Paper [9] presents the smlarty measure n detal. smlarty ( D, Q) = Pr( D Q < ) () D, Q are two feature vectors B. System dagram In our method, we assume that the uncertanty s small [0], therefore the dsturbed objects roughly keep the orgnal dstrbuton. Thus, we can use the feature vector means to construct the ndex, and attach uncertan nformaton to the data entres. At ths step, an ndex whch handles uncertanty s constructed, as shown by the dash-lne box n Fg.. Index whch handles uncertanty Uncertanty s attached to each data entry Feature vectors Feature vector means are extracted Indexng constructon Get nearest node(s) Gather all data entres as the canddate set Refne the query result wthn canddate set based on the smlarty measure NN search results Fg.. Index structure and NN search procedure. Query When a query comes n, the nearest node(s) n the ndex s found. All the data entres belongng to these nodes (wth uncertanty) are gathered as the canddate set. Ths process s called flter. In the refne step, the smlarty between the query and each data entry n the canddate set s calculated and sorted. The entres correspondng to the largest smlarty are the query result. Ths s called nearest neghbors (NN) search. Here, we are only nterested n NN because t s the bass for other comprehensve queres. The flter step dffers n dfferent ndex constructons. We construct two ndex systems: (a) Optmzed Gaussan mxture herarchy () based on Gaussan mxture model, and (b) Two R-tree varants called uncertan R-tree () and. These are explaned n part C and D n ths secton. C. Optmzed Gaussan Mxture Herarchy () We assume all the objects n the n dmensonal feature space follow some dstrbuton. From the probablty theory, any dstrbuton can be approxmated by a weghted sum of several Gaussan dstrbutons [], as shown n (3), where dstrbuton (x) f x, p s approxmated by C Gaussan dstrbutons: ( ) =,,C. α s the weght for each Gaussan dstrbuton represented by mean vector u and covarance matrx Σ : C C p ( x) = α f ( x) = α N( u, Σ ) (3) Fgueredo and Jan [] proposed a varant of Expectaton Maxmzaton (EM) algorthm to automatcally fnd the number of clusters and to perform clusterng. We use ths algorthm to get the Gaussan components. It nvolves the followng steps: In the frst step, the entre dataset s clustered nto several groups. In the subsequent steps, a bottom-up bnary tree s bult based on these Gaussan components. Snce every leaf node s represented by a Gaussan dstrbuton, we represent every nner node by the Gaussan mxture of all ts offsprng leaves. When a query comes n, we start from the root node, calculate the query s smlarty wth the left chld and the rght chld, and then select the branch wth hgher smlarty. Ths process s repeated untl a leaf node s reached. Then all the data entres belongng to ths leaf are gathered as the canddate set. If there are not enough data entres n the canddate set, all data belongng to the sblng of ths leaf node are added. Refnement s performed wthn the canddate set to get the NN search result. Detal explanaton of tree constructon and NN search algorthm are gven n [9]. D. Uncertan R-tree The uncertan R-tree constructon s based on all the feature vector means and uncertan nformaton s attached to each data entry. To support NN searches, two flter strateges: and are developed. ) : Nearest leaves are returned, even f they do not belong to the same parents. The number of leaves s decded by the requred canddate set sze. All data entres of these leaf nodes are gathered together as the canddate set for our uncertan query. ) : The nearest leaf s found, then ts ancestors are backtracked untl the one that has enough data entry
3 offsprng for refnement s met. All the data entres of these offsprng construct the canddate set. The canddate set n both and s refned to get the NN for the query. and have the same canddate set extracton strategy, so only ther comparson s far, whereas wll theoretcally acheve hgher precson. The comparson n the next secton wll testfy ths result. III. EXPERIMENTAL RESULTS A. Dataset and uncertanty assgnment We take the TIGER/Lne southern Calforna landmark pont dataset n the experment, as shown n Fg. and Fg. 3. It contans 8703 D coordnates (longtude and lattude n degrees) and ts precson (uncertanty) s 67 feet [0], whch s Uncertanty s added as a D Gaussan nose to each pont, as shown n (4): Nose ~ N [ 0,0] T δ x, 0 0 δ y (4) for dfferent uncertanty cases. There s a parameter defnng the mnmum canddate set sze, called MCS_sze. ~5 nearest neghbor(s) are returned as the result to the query. B. Comparson of and uncertan R-tree The NN search s made on,, and plan R-tree. Ther performance measures are precson, I/O cost and CPU cost. The experments are performed under two dfferent uncertantes (0.0005, ) and two MCS_szes (40, 60). The program s wrtten n C++ and the system confguraton s as below: System: Sun Mcrosystems sun4u OS: Solars.8 Memory: 048MB ) Precson: # of correct results actually returned Precson = # of results should be returned Precson 5 Fg.. Countes n Southern Calforna. Fg. 4. σ= , MCS_sze = 40. Precson Fg. 3. Landmarks n Southern Calforna. In all the experments, the test data are fxed (the orgnal data) and the tranng data are nosy, whereδ x, δ for each y pont are randomly selected from [0, ] or [0, ] 5 Fg. 5. σ= , MCS sze = 60. 3
4 σ = σ = Precson Precson Fg. 6. σ= 0.005, MCS_sze = 40. Fg. 8. precson on dfferent σ. 6 4 σ = σ = Precson 5 Fg. 7. σ= 0.005, MCS_sze = 60. Precson Fg. 9. precson on dfferent σ. From Fg. 4 - Fg. 7, we can make the followng observatons: (a) always gves the best performance, followed by and. As mentoned n Secton II.D (Uncertan R- tree), and have the same flter strategy. (b) has hgher precson performance than, especally when MCS_sze s larger. So Gaussan Mxture model s more approprate than MBRs n PDF ndexng. (c) Plan R-tree gves the worst precson and t s not acceptable, so n the followng comparsons, plan R-tree s removed. From Fg. 8 and Fg. 9 we can see that when the uncertanty ncreases from to (MCS_sze = 60), the precson performance of degrades much more than that of (5% vs. 0.5%). So s more stable than uncertan R-tree. ) I/O cost -- the average page read/wrte for -NN query sgma = 0.005, MCS_sze = 40 sgma = 0.005, MCS_sze = 60 sgma = , MCS_sze = 40 sgma = , MCS_sze = 60 Fg. 0. I/O cost comparson among all ndces. As shown n Fg. 0, and need more I/O cost than. The tree has more levels than R-tree (7 vs. 4). The more I/O cost of comes from the back trackng, but t s stll comparable wth. 4
5 ) CPU cost -- average tme (second) for -NN query. sgma = 0.005, MCS_sze= 40 sgma = 0.005, MCS_sze = 60 sgma = , MCS_sze = 40 sgma = , MCS_sze = 60 Fg.. CPU cost comparson among all ndces. Fg. ndcates that and are comparable on tme complexty and they are more effcent than, ths s because the back trackng of s tme consumng. From the three comparsons above, we see that when the query s fxed, s the best n general, followed by and. Plan R-tree s not acceptable wth respect to precson performance. But as mentoned n Secton I, a complete system should be able to handle both fxed and uncertan queres. Thus, the choce of ndex structure depends on the applcaton. If no uncertan query s asked, s the best choce, otherwse only s sutable. IV. CONCLUSIONS Uncertanty n spatal databases s gettng more and more attenton, but most of the exstng approaches are based on relatonal models usng fuzzy set theory. Ths method s only sutable for handlng uncertanty n symbolc level. In order to support uncertantes at numerc and exstence levels, we proposed a new uncertanty model and new ndexng schemes. In ths paper, we represented uncertan objects wth PDFs and constructed an optmzed Gaussan mxture herarchy based on Gaussan mxture model and an uncertan R-tree. After a comprehensve comparson based on NN search precson, I/O cost and CPU cost, uncertan R-tree () s the best for fxed queres. But s sutable for both certan and uncertan queres. Moreover, the s not sutable for only spatal databases, but also for other multdmensonal ndex applcatons lke content based mage and vdeo retreval. REFERENCES [] Rgaus, P., M. Scholl, and A. Vosard, Spatal Databases: wth Applcaton to GIS. San Francsco, Calforna: Morgan aufmann, 00. [] Internatonal Symposum on Spatal Accuracy Assessment n Natural Resources and Envronmental Scences. ault.html [3] Zanolo, C., et al., Introducton to Advanced Database Systems: Morgan-aufmann, 997. [4] Schneder, M. Uncertanty Management for Spatal Data n Databases: Fuzzy Spatal Data Types. n 6th Int. Symposum on Advances n Spatal Databases (SSD). 999: Sprnger Verlag. [5] Robnson, V.B., A Perspectve on Managng Uncertanty n Geographc Informaton System wth Fuzzy Sets. IEEE Transactons n Geographc Informaton System, (): p. -5. [6] Guttman, A. R-trees: A Dynamc Index Structure for Spatal Searchng. n SIGMOD Boston, MA,. [7] Greene, D. An Implementaton and Performance Analyss of Spatal Data Access Methods. n Proceedngs of the Ffth Internatonal Conference on Data Engneerng. 989: IEEE Computer Socety Washngton, DC, USA. [8] Beckmann, N., et al. The R*-tree: an effcent and robust access method for ponts and rectangles. n ACM SIGMOD Internatonal Conference on Management of Data Atlantc Cty, NJ. [9] Bhanu, B., R. L, C. Ravshankar, M. urth, and J. N. Indexng Structure for Handlng Uncertan Spatal Database. n 6 th Internatonal Symposum on Spatal Accuracy Assessment n Natural Resources and Envronmental Scences Portland, Mane, USA. [0] Brown, R.H. and E. Ehrlch, TIGER/Lne(TM) Fles, : Washngton, D. C. (03/30/04) [] Duda, R.O., Hart, P.E., and D.G. Stork, Pattern Classfcaton. New York: A Wley-Interscence Publcaton, 000. [] Fgueredo, M.A. and A. Jan, Unsupervsed Learnng of Fnte Mxture Models. IEEE Transactons on Pattern Analyss and Machne Intellgence, 00. 4(3): p
Semi-supervised Classification with Active Query Selection
Sem-supervsed Classfcaton wth Actve Query Selecton Jao Wang and Swe Luo School of Computer and Informaton Technology, Beng Jaotong Unversty, Beng 00044, Chna Wangjao088@63.com Abstract. Labeled samples
More informationRegularized Discriminant Analysis for Face Recognition
1 Regularzed Dscrmnant Analyss for Face Recognton Itz Pma, Mayer Aladem Department of Electrcal and Computer Engneerng, Ben-Guron Unversty of the Negev P.O.Box 653, Beer-Sheva, 845, Israel. Abstract Ths
More informationAutomatic Object Trajectory- Based Motion Recognition Using Gaussian Mixture Models
Automatc Object Trajectory- Based Moton Recognton Usng Gaussan Mxture Models Fasal I. Bashr, Ashfaq A. Khokhar, Dan Schonfeld Electrcal and Computer Engneerng, Unversty of Illnos at Chcago. Chcago, IL,
More informationSupport Vector Machines. Vibhav Gogate The University of Texas at dallas
Support Vector Machnes Vbhav Gogate he Unversty of exas at dallas What We have Learned So Far? 1. Decson rees. Naïve Bayes 3. Lnear Regresson 4. Logstc Regresson 5. Perceptron 6. Neural networks 7. K-Nearest
More informationSpatial Statistics and Analysis Methods (for GEOG 104 class).
Spatal Statstcs and Analyss Methods (for GEOG 104 class). Provded by Dr. An L, San Dego State Unversty. 1 Ponts Types of spatal data Pont pattern analyss (PPA; such as nearest neghbor dstance, quadrat
More informationGEMINI GEneric Multimedia INdexIng
GEMINI GEnerc Multmeda INdexIng Last lecture, LSH http://www.mt.edu/~andon/lsh/ Is there another possble soluton? Do we need to perform ANN? 1 GEnerc Multmeda INdexIng dstance measure Sub-pattern Match
More informationA New Scrambling Evaluation Scheme based on Spatial Distribution Entropy and Centroid Difference of Bit-plane
A New Scramblng Evaluaton Scheme based on Spatal Dstrbuton Entropy and Centrod Dfference of Bt-plane Lang Zhao *, Avshek Adhkar Kouch Sakura * * Graduate School of Informaton Scence and Electrcal Engneerng,
More informationAn Improved multiple fractal algorithm
Advanced Scence and Technology Letters Vol.31 (MulGraB 213), pp.184-188 http://dx.do.org/1.1427/astl.213.31.41 An Improved multple fractal algorthm Yun Ln, Xaochu Xu, Jnfeng Pang College of Informaton
More informationLecture 4: Universal Hash Functions/Streaming Cont d
CSE 5: Desgn and Analyss of Algorthms I Sprng 06 Lecture 4: Unversal Hash Functons/Streamng Cont d Lecturer: Shayan Oves Gharan Aprl 6th Scrbe: Jacob Schreber Dsclamer: These notes have not been subjected
More informationGaussian Mixture Models
Lab Gaussan Mxture Models Lab Objectve: Understand the formulaton of Gaussan Mxture Models (GMMs) and how to estmate GMM parameters. You ve already seen GMMs as the observaton dstrbuton n certan contnuous
More informationLecture 10 Support Vector Machines II
Lecture 10 Support Vector Machnes II 22 February 2016 Taylor B. Arnold Yale Statstcs STAT 365/665 1/28 Notes: Problem 3 s posted and due ths upcomng Frday There was an early bug n the fake-test data; fxed
More informationA New Evolutionary Computation Based Approach for Learning Bayesian Network
Avalable onlne at www.scencedrect.com Proceda Engneerng 15 (2011) 4026 4030 Advanced n Control Engneerng and Informaton Scence A New Evolutonary Computaton Based Approach for Learnng Bayesan Network Yungang
More informationAppendix B: Resampling Algorithms
407 Appendx B: Resamplng Algorthms A common problem of all partcle flters s the degeneracy of weghts, whch conssts of the unbounded ncrease of the varance of the mportance weghts ω [ ] of the partcles
More informationA Bayes Algorithm for the Multitask Pattern Recognition Problem Direct Approach
A Bayes Algorthm for the Multtask Pattern Recognton Problem Drect Approach Edward Puchala Wroclaw Unversty of Technology, Char of Systems and Computer etworks, Wybrzeze Wyspanskego 7, 50-370 Wroclaw, Poland
More informationProbability Density Function Estimation by different Methods
EEE 739Q SPRIG 00 COURSE ASSIGMET REPORT Probablty Densty Functon Estmaton by dfferent Methods Vas Chandraant Rayar Abstract The am of the assgnment was to estmate the probablty densty functon (PDF of
More informationModule 3 LOSSY IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur
Module 3 LOSSY IMAGE COMPRESSION SYSTEMS Verson ECE IIT, Kharagpur Lesson 6 Theory of Quantzaton Verson ECE IIT, Kharagpur Instructonal Objectves At the end of ths lesson, the students should be able to:
More informationLecture 12: Classification
Lecture : Classfcaton g Dscrmnant functons g The optmal Bayes classfer g Quadratc classfers g Eucldean and Mahalanobs metrcs g K Nearest Neghbor Classfers Intellgent Sensor Systems Rcardo Guterrez-Osuna
More informationVQ widely used in coding speech, image, and video
at Scalar quantzers are specal cases of vector quantzers (VQ): they are constraned to look at one sample at a tme (memoryless) VQ does not have such constrant better RD perfomance expected Source codng
More informationComparison of the Population Variance Estimators. of 2-Parameter Exponential Distribution Based on. Multiple Criteria Decision Making Method
Appled Mathematcal Scences, Vol. 7, 0, no. 47, 07-0 HIARI Ltd, www.m-hkar.com Comparson of the Populaton Varance Estmators of -Parameter Exponental Dstrbuton Based on Multple Crtera Decson Makng Method
More informationNatural Images, Gaussian Mixtures and Dead Leaves Supplementary Material
Natural Images, Gaussan Mxtures and Dead Leaves Supplementary Materal Danel Zoran Interdscplnary Center for Neural Computaton Hebrew Unversty of Jerusalem Israel http://www.cs.huj.ac.l/ danez Yar Wess
More informationClustering gene expression data & the EM algorithm
CG, Fall 2011-12 Clusterng gene expresson data & the EM algorthm CG 08 Ron Shamr 1 How Gene Expresson Data Looks Entres of the Raw Data matrx: Rato values Absolute values Row = gene s expresson pattern
More informationHomework Assignment 3 Due in class, Thursday October 15
Homework Assgnment 3 Due n class, Thursday October 15 SDS 383C Statstcal Modelng I 1 Rdge regresson and Lasso 1. Get the Prostrate cancer data from http://statweb.stanford.edu/~tbs/elemstatlearn/ datasets/prostate.data.
More informationA Network Intrusion Detection Method Based on Improved K-means Algorithm
Advanced Scence and Technology Letters, pp.429-433 http://dx.do.org/10.14257/astl.2014.53.89 A Network Intruson Detecton Method Based on Improved K-means Algorthm Meng Gao 1,1, Nhong Wang 1, 1 Informaton
More informationParameter Estimation for Dynamic System using Unscented Kalman filter
Parameter Estmaton for Dynamc System usng Unscented Kalman flter Jhoon Seung 1,a, Amr Atya F. 2,b, Alexander G.Parlos 3,c, and Klto Chong 1,4,d* 1 Dvson of Electroncs Engneerng, Chonbuk Natonal Unversty,
More informationMULTISPECTRAL IMAGE CLASSIFICATION USING BACK-PROPAGATION NEURAL NETWORK IN PCA DOMAIN
MULTISPECTRAL IMAGE CLASSIFICATION USING BACK-PROPAGATION NEURAL NETWORK IN PCA DOMAIN S. Chtwong, S. Wtthayapradt, S. Intajag, and F. Cheevasuvt Faculty of Engneerng, Kng Mongkut s Insttute of Technology
More informationU-Pb Geochronology Practical: Background
U-Pb Geochronology Practcal: Background Basc Concepts: accuracy: measure of the dfference between an expermental measurement and the true value precson: measure of the reproducblty of the expermental result
More informationPower law and dimension of the maximum value for belief distribution with the max Deng entropy
Power law and dmenson of the maxmum value for belef dstrbuton wth the max Deng entropy Bngy Kang a, a College of Informaton Engneerng, Northwest A&F Unversty, Yanglng, Shaanx, 712100, Chna. Abstract Deng
More informationProbability Theory (revisited)
Probablty Theory (revsted) Summary Probablty v.s. plausblty Random varables Smulaton of Random Experments Challenge The alarm of a shop rang. Soon afterwards, a man was seen runnng n the street, persecuted
More informationKernel Methods and SVMs Extension
Kernel Methods and SVMs Extenson The purpose of ths document s to revew materal covered n Machne Learnng 1 Supervsed Learnng regardng support vector machnes (SVMs). Ths document also provdes a general
More informationProblem Set 9 Solutions
Desgn and Analyss of Algorthms May 4, 2015 Massachusetts Insttute of Technology 6.046J/18.410J Profs. Erk Demane, Srn Devadas, and Nancy Lynch Problem Set 9 Solutons Problem Set 9 Solutons Ths problem
More informationAggregation of Social Networks by Divisive Clustering Method
ggregaton of Socal Networks by Dvsve Clusterng Method mne Louat and Yves Lechaveller INRI Pars-Rocquencourt Rocquencourt, France {lzennyr.da_slva, Yves.Lechevaller, Fabrce.Ross}@nra.fr HCSD Beng October
More informationOn an Extension of Stochastic Approximation EM Algorithm for Incomplete Data Problems. Vahid Tadayon 1
On an Extenson of Stochastc Approxmaton EM Algorthm for Incomplete Data Problems Vahd Tadayon Abstract: The Stochastc Approxmaton EM (SAEM algorthm, a varant stochastc approxmaton of EM, s a versatle tool
More informationLinear Classification, SVMs and Nearest Neighbors
1 CSE 473 Lecture 25 (Chapter 18) Lnear Classfcaton, SVMs and Nearest Neghbors CSE AI faculty + Chrs Bshop, Dan Klen, Stuart Russell, Andrew Moore Motvaton: Face Detecton How do we buld a classfer to dstngush
More informationInductance Calculation for Conductors of Arbitrary Shape
CRYO/02/028 Aprl 5, 2002 Inductance Calculaton for Conductors of Arbtrary Shape L. Bottura Dstrbuton: Internal Summary In ths note we descrbe a method for the numercal calculaton of nductances among conductors
More informationADVANCED MACHINE LEARNING ADVANCED MACHINE LEARNING
1 ADVANCED ACHINE LEARNING ADVANCED ACHINE LEARNING Non-lnear regresson technques 2 ADVANCED ACHINE LEARNING Regresson: Prncple N ap N-dm. nput x to a contnuous output y. Learn a functon of the type: N
More information1 The Mistake Bound Model
5-850: Advanced Algorthms CMU, Sprng 07 Lecture #: Onlne Learnng and Multplcatve Weghts February 7, 07 Lecturer: Anupam Gupta Scrbe: Bryan Lee,Albert Gu, Eugene Cho he Mstake Bound Model Suppose there
More informationχ x B E (c) Figure 2.1.1: (a) a material particle in a body, (b) a place in space, (c) a configuration of the body
Secton.. Moton.. The Materal Body and Moton hyscal materals n the real world are modeled usng an abstract mathematcal entty called a body. Ths body conssts of an nfnte number of materal partcles. Shown
More informationUsing T.O.M to Estimate Parameter of distributions that have not Single Exponential Family
IOSR Journal of Mathematcs IOSR-JM) ISSN: 2278-5728. Volume 3, Issue 3 Sep-Oct. 202), PP 44-48 www.osrjournals.org Usng T.O.M to Estmate Parameter of dstrbutons that have not Sngle Exponental Famly Jubran
More informationHopfield networks and Boltzmann machines. Geoffrey Hinton et al. Presented by Tambet Matiisen
Hopfeld networks and Boltzmann machnes Geoffrey Hnton et al. Presented by Tambet Matsen 18.11.2014 Hopfeld network Bnary unts Symmetrcal connectons http://www.nnwj.de/hopfeld-net.html Energy functon The
More informationLecture Notes on Linear Regression
Lecture Notes on Lnear Regresson Feng L fl@sdueducn Shandong Unversty, Chna Lnear Regresson Problem In regresson problem, we am at predct a contnuous target value gven an nput feature vector We assume
More informationLecture 3: Dual problems and Kernels
Lecture 3: Dual problems and Kernels C4B Machne Learnng Hlary 211 A. Zsserman Prmal and dual forms Lnear separablty revsted Feature mappng Kernels for SVMs Kernel trck requrements radal bass functons SVM
More informationPop-Click Noise Detection Using Inter-Frame Correlation for Improved Portable Auditory Sensing
Advanced Scence and Technology Letters, pp.164-168 http://dx.do.org/10.14257/astl.2013 Pop-Clc Nose Detecton Usng Inter-Frame Correlaton for Improved Portable Audtory Sensng Dong Yun Lee, Kwang Myung Jeon,
More informationIntroduction to Information Theory, Data Compression,
Introducton to Informaton Theory, Data Compresson, Codng Mehd Ibm Brahm, Laura Mnkova Aprl 5, 208 Ths s the augmented transcrpt of a lecture gven by Luc Devroye on the 3th of March 208 for a Data Structures
More informationGeneralized Linear Methods
Generalzed Lnear Methods 1 Introducton In the Ensemble Methods the general dea s that usng a combnaton of several weak learner one could make a better learner. More formally, assume that we have a set
More informationDiscretization of Continuous Attributes in Rough Set Theory and Its Application*
Dscretzaton of Contnuous Attrbutes n Rough Set Theory and Its Applcaton* Gexang Zhang 1,2, Lazhao Hu 1, and Wedong Jn 2 1 Natonal EW Laboratory, Chengdu 610036 Schuan, Chna dylan7237@sna.com 2 School of
More informationCS 468 Lecture 16: Isometry Invariance and Spectral Techniques
CS 468 Lecture 16: Isometry Invarance and Spectral Technques Justn Solomon Scrbe: Evan Gawlk Introducton. In geometry processng, t s often desrable to characterze the shape of an object n a manner that
More informationANSWERS. Problem 1. and the moment generating function (mgf) by. defined for any real t. Use this to show that E( U) var( U)
Econ 413 Exam 13 H ANSWERS Settet er nndelt 9 deloppgaver, A,B,C, som alle anbefales å telle lkt for å gøre det ltt lettere å stå. Svar er gtt . Unfortunately, there s a prntng error n the hnt of
More informationImprovement of Histogram Equalization for Minimum Mean Brightness Error
Proceedngs of the 7 WSEAS Int. Conference on Crcuts, Systems, Sgnal and elecommuncatons, Gold Coast, Australa, January 7-9, 7 3 Improvement of Hstogram Equalzaton for Mnmum Mean Brghtness Error AAPOG PHAHUA*,
More informationNotes on Frequency Estimation in Data Streams
Notes on Frequency Estmaton n Data Streams In (one of) the data streamng model(s), the data s a sequence of arrvals a 1, a 2,..., a m of the form a j = (, v) where s the dentty of the tem and belongs to
More informationExcess Error, Approximation Error, and Estimation Error
E0 370 Statstcal Learnng Theory Lecture 10 Sep 15, 011 Excess Error, Approxaton Error, and Estaton Error Lecturer: Shvan Agarwal Scrbe: Shvan Agarwal 1 Introducton So far, we have consdered the fnte saple
More informationSTATS 306B: Unsupervised Learning Spring Lecture 10 April 30
STATS 306B: Unsupervsed Learnng Sprng 2014 Lecture 10 Aprl 30 Lecturer: Lester Mackey Scrbe: Joey Arthur, Rakesh Achanta 10.1 Factor Analyss 10.1.1 Recap Recall the factor analyss (FA) model for lnear
More informationBACKGROUND SUBTRACTION WITH EIGEN BACKGROUND METHODS USING MATLAB
BACKGROUND SUBTRACTION WITH EIGEN BACKGROUND METHODS USING MATLAB 1 Ilmyat Sar 2 Nola Marna 1 Pusat Stud Komputas Matematka, Unverstas Gunadarma e-mal: lmyat@staff.gunadarma.ac.d 2 Pusat Stud Komputas
More informationEnsemble Methods: Boosting
Ensemble Methods: Boostng Ncholas Ruozz Unversty of Texas at Dallas Based on the sldes of Vbhav Gogate and Rob Schapre Last Tme Varance reducton va baggng Generate new tranng data sets by samplng wth replacement
More informationLinear Regression Analysis: Terminology and Notation
ECON 35* -- Secton : Basc Concepts of Regresson Analyss (Page ) Lnear Regresson Analyss: Termnology and Notaton Consder the generc verson of the smple (two-varable) lnear regresson model. It s represented
More informationCOMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS
Avalable onlne at http://sck.org J. Math. Comput. Sc. 3 (3), No., 6-3 ISSN: 97-537 COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS
More informationRBF Neural Network Model Training by Unscented Kalman Filter and Its Application in Mechanical Fault Diagnosis
Appled Mechancs and Materals Submtted: 24-6-2 ISSN: 662-7482, Vols. 62-65, pp 2383-2386 Accepted: 24-6- do:.428/www.scentfc.net/amm.62-65.2383 Onlne: 24-8- 24 rans ech Publcatons, Swtzerland RBF Neural
More informationNumerical Heat and Mass Transfer
Master degree n Mechancal Engneerng Numercal Heat and Mass Transfer 06-Fnte-Dfference Method (One-dmensonal, steady state heat conducton) Fausto Arpno f.arpno@uncas.t Introducton Why we use models and
More informationLarge-Margin HMM Estimation for Speech Recognition
Large-Margn HMM Estmaton for Speech Recognton Prof. Hu Jang Department of Computer Scence and Engneerng York Unversty, Toronto, Ont. M3J 1P3, CANADA Emal: hj@cs.yorku.ca Ths s a jont work wth Chao-Jun
More informationEcon107 Applied Econometrics Topic 3: Classical Model (Studenmund, Chapter 4)
I. Classcal Assumptons Econ7 Appled Econometrcs Topc 3: Classcal Model (Studenmund, Chapter 4) We have defned OLS and studed some algebrac propertes of OLS. In ths topc we wll study statstcal propertes
More informationInternational Journal of Mathematical Archive-3(3), 2012, Page: Available online through ISSN
Internatonal Journal of Mathematcal Archve-3(3), 2012, Page: 1136-1140 Avalable onlne through www.ma.nfo ISSN 2229 5046 ARITHMETIC OPERATIONS OF FOCAL ELEMENTS AND THEIR CORRESPONDING BASIC PROBABILITY
More informationMicrowave Diversity Imaging Compression Using Bioinspired
Mcrowave Dversty Imagng Compresson Usng Bonspred Neural Networks Youwe Yuan 1, Yong L 1, Wele Xu 1, Janghong Yu * 1 School of Computer Scence and Technology, Hangzhou Danz Unversty, Hangzhou, Zhejang,
More informationCONTRAST ENHANCEMENT FOR MIMIMUM MEAN BRIGHTNESS ERROR FROM HISTOGRAM PARTITIONING INTRODUCTION
CONTRAST ENHANCEMENT FOR MIMIMUM MEAN BRIGHTNESS ERROR FROM HISTOGRAM PARTITIONING N. Phanthuna 1,2, F. Cheevasuvt 2 and S. Chtwong 2 1 Department of Electrcal Engneerng, Faculty of Engneerng Rajamangala
More informationCSci 6974 and ECSE 6966 Math. Tech. for Vision, Graphics and Robotics Lecture 21, April 17, 2006 Estimating A Plane Homography
CSc 6974 and ECSE 6966 Math. Tech. for Vson, Graphcs and Robotcs Lecture 21, Aprl 17, 2006 Estmatng A Plane Homography Overvew We contnue wth a dscusson of the major ssues, usng estmaton of plane projectve
More informationISSN: ISO 9001:2008 Certified International Journal of Engineering and Innovative Technology (IJEIT) Volume 3, Issue 1, July 2013
ISSN: 2277-375 Constructon of Trend Free Run Orders for Orthogonal rrays Usng Codes bstract: Sometmes when the expermental runs are carred out n a tme order sequence, the response can depend on the run
More informationModule 9. Lecture 6. Duality in Assignment Problems
Module 9 1 Lecture 6 Dualty n Assgnment Problems In ths lecture we attempt to answer few other mportant questons posed n earler lecture for (AP) and see how some of them can be explaned through the concept
More informationSpeeding up Computation of Scalar Multiplication in Elliptic Curve Cryptosystem
H.K. Pathak et. al. / (IJCSE) Internatonal Journal on Computer Scence and Engneerng Speedng up Computaton of Scalar Multplcaton n Ellptc Curve Cryptosystem H. K. Pathak Manju Sangh S.o.S n Computer scence
More informationDesign and Optimization of Fuzzy Controller for Inverse Pendulum System Using Genetic Algorithm
Desgn and Optmzaton of Fuzzy Controller for Inverse Pendulum System Usng Genetc Algorthm H. Mehraban A. Ashoor Unversty of Tehran Unversty of Tehran h.mehraban@ece.ut.ac.r a.ashoor@ece.ut.ac.r Abstract:
More informationChapter 8 Indicator Variables
Chapter 8 Indcator Varables In general, e explanatory varables n any regresson analyss are assumed to be quanttatve n nature. For example, e varables lke temperature, dstance, age etc. are quanttatve n
More informationFinite Mixture Models and Expectation Maximization. Most slides are from: Dr. Mario Figueiredo, Dr. Anil Jain and Dr. Rong Jin
Fnte Mxture Models and Expectaton Maxmzaton Most sldes are from: Dr. Maro Fgueredo, Dr. Anl Jan and Dr. Rong Jn Recall: The Supervsed Learnng Problem Gven a set of n samples X {(x, y )},,,n Chapter 3 of
More informationA Fast Fractal Image Compression Algorithm Using Predefined Values for Contrast Scaling
Proceedngs of the World Congress on Engneerng and Computer Scence 007 WCECS 007, October 4-6, 007, San Francsco, USA A Fast Fractal Image Compresson Algorthm Usng Predefned Values for Contrast Scalng H.
More informationEM and Structure Learning
EM and Structure Learnng Le Song Machne Learnng II: Advanced Topcs CSE 8803ML, Sprng 2012 Partally observed graphcal models Mxture Models N(μ 1, Σ 1 ) Z X N N(μ 2, Σ 2 ) 2 Gaussan mxture model Consder
More informationThe Order Relation and Trace Inequalities for. Hermitian Operators
Internatonal Mathematcal Forum, Vol 3, 08, no, 507-57 HIKARI Ltd, wwwm-hkarcom https://doorg/0988/mf088055 The Order Relaton and Trace Inequaltes for Hermtan Operators Y Huang School of Informaton Scence
More informationSociété de Calcul Mathématique SA
Socété de Calcul Mathématque SA Outls d'ade à la décson Tools for decson help Probablstc Studes: Normalzng the Hstograms Bernard Beauzamy December, 202 I. General constructon of the hstogram Any probablstc
More informationNeuro-Adaptive Design - I:
Lecture 36 Neuro-Adaptve Desgn - I: A Robustfyng ool for Dynamc Inverson Desgn Dr. Radhakant Padh Asst. Professor Dept. of Aerospace Engneerng Indan Insttute of Scence - Bangalore Motvaton Perfect system
More informationA PROBABILITY-DRIVEN SEARCH ALGORITHM FOR SOLVING MULTI-OBJECTIVE OPTIMIZATION PROBLEMS
HCMC Unversty of Pedagogy Thong Nguyen Huu et al. A PROBABILITY-DRIVEN SEARCH ALGORITHM FOR SOLVING MULTI-OBJECTIVE OPTIMIZATION PROBLEMS Thong Nguyen Huu and Hao Tran Van Department of mathematcs-nformaton,
More informationOnline Classification: Perceptron and Winnow
E0 370 Statstcal Learnng Theory Lecture 18 Nov 8, 011 Onlne Classfcaton: Perceptron and Wnnow Lecturer: Shvan Agarwal Scrbe: Shvan Agarwal 1 Introducton In ths lecture we wll start to study the onlne learnng
More informationLecture 4: November 17, Part 1 Single Buffer Management
Lecturer: Ad Rosén Algorthms for the anagement of Networs Fall 2003-2004 Lecture 4: November 7, 2003 Scrbe: Guy Grebla Part Sngle Buffer anagement In the prevous lecture we taled about the Combned Input
More informationA Particle Filter Algorithm based on Mixing of Prior probability density and UKF as Generate Importance Function
Advanced Scence and Technology Letters, pp.83-87 http://dx.do.org/10.14257/astl.2014.53.20 A Partcle Flter Algorthm based on Mxng of Pror probablty densty and UKF as Generate Importance Functon Lu Lu 1,1,
More informationCalculation of time complexity (3%)
Problem 1. (30%) Calculaton of tme complexty (3%) Gven n ctes, usng exhaust search to see every result takes O(n!). Calculaton of tme needed to solve the problem (2%) 40 ctes:40! dfferent tours 40 add
More informationFuzzy Boundaries of Sample Selection Model
Proceedngs of the 9th WSES Internatonal Conference on ppled Mathematcs, Istanbul, Turkey, May 7-9, 006 (pp309-34) Fuzzy Boundares of Sample Selecton Model L. MUHMD SFIIH, NTON BDULBSH KMIL, M. T. BU OSMN
More informationFinding Dense Subgraphs in G(n, 1/2)
Fndng Dense Subgraphs n Gn, 1/ Atsh Das Sarma 1, Amt Deshpande, and Rav Kannan 1 Georga Insttute of Technology,atsh@cc.gatech.edu Mcrosoft Research-Bangalore,amtdesh,annan@mcrosoft.com Abstract. Fndng
More informationStructure and Drive Paul A. Jensen Copyright July 20, 2003
Structure and Drve Paul A. Jensen Copyrght July 20, 2003 A system s made up of several operatons wth flow passng between them. The structure of the system descrbes the flow paths from nputs to outputs.
More informationMarkov Chain Monte Carlo Lecture 6
where (x 1,..., x N ) X N, N s called the populaton sze, f(x) f (x) for at least one {1, 2,..., N}, and those dfferent from f(x) are called the tral dstrbutons n terms of mportance samplng. Dfferent ways
More informationNon-linear Canonical Correlation Analysis Using a RBF Network
ESANN' proceedngs - European Smposum on Artfcal Neural Networks Bruges (Belgum), 4-6 Aprl, d-sde publ., ISBN -97--, pp. 57-5 Non-lnear Canoncal Correlaton Analss Usng a RBF Network Sukhbnder Kumar, Elane
More informationHidden Markov Models & The Multivariate Gaussian (10/26/04)
CS281A/Stat241A: Statstcal Learnng Theory Hdden Markov Models & The Multvarate Gaussan (10/26/04) Lecturer: Mchael I. Jordan Scrbes: Jonathan W. Hu 1 Hdden Markov Models As a bref revew, hdden Markov models
More informationMathematical Preparations
1 Introducton Mathematcal Preparatons The theory of relatvty was developed to explan experments whch studed the propagaton of electromagnetc radaton n movng coordnate systems. Wthn expermental error the
More information4DVAR, according to the name, is a four-dimensional variational method.
4D-Varatonal Data Assmlaton (4D-Var) 4DVAR, accordng to the name, s a four-dmensonal varatonal method. 4D-Var s actually a drect generalzaton of 3D-Var to handle observatons that are dstrbuted n tme. The
More informationGeneral theory of fuzzy connectedness segmentations: reconciliation of two tracks of FC theory
General theory of fuzzy connectedness segmentatons: reconclaton of two tracks of FC theory Krzysztof Chrs Ceselsk Department of Mathematcs, West Vrgna Unversty and MIPG, Department of Radology, Unversty
More informationUsing Immune Genetic Algorithm to Optimize BP Neural Network and Its Application Peng-fei LIU1,Qun-tai SHEN1 and Jun ZHI2,*
Advances n Computer Scence Research (ACRS), volume 54 Internatonal Conference on Computer Networks and Communcaton Technology (CNCT206) Usng Immune Genetc Algorthm to Optmze BP Neural Network and Its Applcaton
More informationEstimating the Fundamental Matrix by Transforming Image Points in Projective Space 1
Estmatng the Fundamental Matrx by Transformng Image Ponts n Projectve Space 1 Zhengyou Zhang and Charles Loop Mcrosoft Research, One Mcrosoft Way, Redmond, WA 98052, USA E-mal: fzhang,cloopg@mcrosoft.com
More informationAn efficient algorithm for multivariate Maclaurin Newton transformation
Annales UMCS Informatca AI VIII, 2 2008) 5 14 DOI: 10.2478/v10065-008-0020-6 An effcent algorthm for multvarate Maclaurn Newton transformaton Joanna Kapusta Insttute of Mathematcs and Computer Scence,
More informationThe Expectation-Maximization Algorithm
The Expectaton-Maxmaton Algorthm Charles Elan elan@cs.ucsd.edu November 16, 2007 Ths chapter explans the EM algorthm at multple levels of generalty. Secton 1 gves the standard hgh-level verson of the algorthm.
More informationLearning from Data 1 Naive Bayes
Learnng from Data 1 Nave Bayes Davd Barber dbarber@anc.ed.ac.uk course page : http://anc.ed.ac.uk/ dbarber/lfd1/lfd1.html c Davd Barber 2001, 2002 1 Learnng from Data 1 : c Davd Barber 2001,2002 2 1 Why
More informationOne-sided finite-difference approximations suitable for use with Richardson extrapolation
Journal of Computatonal Physcs 219 (2006) 13 20 Short note One-sded fnte-dfference approxmatons sutable for use wth Rchardson extrapolaton Kumar Rahul, S.N. Bhattacharyya * Department of Mechancal Engneerng,
More informationRelevance Vector Machines Explained
October 19, 2010 Relevance Vector Machnes Explaned Trstan Fletcher www.cs.ucl.ac.uk/staff/t.fletcher/ Introducton Ths document has been wrtten n an attempt to make Tppng s [1] Relevance Vector Machnes
More informationFeature Selection & Dynamic Tracking F&P Textbook New: Ch 11, Old: Ch 17 Guido Gerig CS 6320, Spring 2013
Feature Selecton & Dynamc Trackng F&P Textbook New: Ch 11, Old: Ch 17 Gudo Gerg CS 6320, Sprng 2013 Credts: Materal Greg Welch & Gary Bshop, UNC Chapel Hll, some sldes modfed from J.M. Frahm/ M. Pollefeys,
More informationINF 5860 Machine learning for image classification. Lecture 3 : Image classification and regression part II Anne Solberg January 31, 2018
INF 5860 Machne learnng for mage classfcaton Lecture 3 : Image classfcaton and regresson part II Anne Solberg January 3, 08 Today s topcs Multclass logstc regresson and softma Regularzaton Image classfcaton
More informationSupporting Information
Supportng Informaton The neural network f n Eq. 1 s gven by: f x l = ReLU W atom x l + b atom, 2 where ReLU s the element-wse rectfed lnear unt, 21.e., ReLUx = max0, x, W atom R d d s the weght matrx to
More informationBoostrapaggregating (Bagging)
Boostrapaggregatng (Baggng) An ensemble meta-algorthm desgned to mprove the stablty and accuracy of machne learnng algorthms Can be used n both regresson and classfcaton Reduces varance and helps to avod
More informationGrover s Algorithm + Quantum Zeno Effect + Vaidman
Grover s Algorthm + Quantum Zeno Effect + Vadman CS 294-2 Bomb 10/12/04 Fall 2004 Lecture 11 Grover s algorthm Recall that Grover s algorthm for searchng over a space of sze wors as follows: consder the
More information