The Degrees of Freedom of Partial Least Squares Regression
|
|
- Donald Reeves
- 6 years ago
- Views:
Transcription
1 The Degrees of Freedom of Partial Least Squares Regression Dr. Nicole Krämer TU München 5th ESSEC-SUPELEC Research Workshop May 20, 2011
2 My talk is about the statistical analysis of Partial Least Squares Regression. 1. intrinsic complexity of PLS 2. comparison of regression methods 3. model selection 4. variable selection based on confidence intervals The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 2 / 27
3 Example: Near Infrared Spectroscopy Predict Y the percentage of water in meat based on X its near infra red spectrum. unknown linear relationship f (x) = β 0 + β, x = β 0 + We observe p β j x (j). j=1 y i f (x i ), i = 1,..., n. x 1 centered data X =. Rn p y = (y 1,..., y n) R n. x n The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 3 / 27
4 Partial Least Squares in one Slide Partial Least Squares (PLS) = 1. supervised dimensionality reduction 2. + least squares regression n p X Least Squares Regression 1 y n n supervised dimensionality reduction T m Partial Least Squares Regression The PLS components T have maximal covariance to the response variable y. The m p components T are used as new predictor variables in a least-squares fit. The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 4 / 27
5 Partial Least Squares Algorithm p 1 Partial Least Squares (PLS) = 1. supervised dimensionality reduction 2. + least squares regression PLS components T have maximal covariance to y. n n X supervised dimensionality reduction T Least Squares Regression Partial Least Squares Regression y n m Algorithm (NIPALS) X 1 = X. For i = 1,..., m model parameter 1. w i X i y maximize covariance cov(x i w i,y) w w = w i X i y w w 2. t i X i w i latent component 3. X i+1 = X i t i t i X i enforce orthogonality β m = W (T XW ) 1 T y Return T = (t 1,..., t m) and W = (w 1,..., w m). The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 5 / 27
6 Degrees of Freedom: Why are they important? Degrees of Freedom (DoF) 1. capture the intrinsic complexity of a regression method. Y i = f (x i ) + ε i, ε i N ( 0, σ 2), 2. are used for model selection. test error = training error + complexity(dof) Examples: Bayesian Information Criterion BIC = y ŷ 2 n + log(n)var (ε) DoF n Akaike Information Criterion, Minimum Description Length,... The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 6 / 27
7 Definition: Degrees of Freedom Assumption: The regression method is linear, i.e. ŷ = Hy, with H independent of y. Degrees of Freedom The Degrees of Freedom of a linear fitting method are DoF = trace (H). Examples Principal Components Regression with m components DoF(m) = 1 + m Ridge Regression, Smoothing splines,... The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 7 / 27
8 Naive Degrees of Freedom for PLS? Recall: PLS is not linear in y p 1 ( ) 1 ŷ m = y + T T T T y. }{{} =: H depends on y n n X supervised dimensionality reduction T Least Squares Regression Partial Least Squares Regression y n m Degrees of Freedom 1 + trace( H) = 1 + m. If we ignore the nonlinearity, we obtain DoF naive (m) = 1 + m. The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 8 / 27
9 Degrees of Freedom for PLS (K. & Braun, 2007; K. & Sugiyama, 2011) The generalized Degrees of Freedom of a regression method are ( )] ŷ DoF = E Y [trace y This coincides with the previous definition if the method is linear. Proposition An unbiased estimate of the Degrees of Freedom of PLS is ( ) ŷm DoF(m) = trace. y We need to compute (the trace of) the first derivative. The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 9 / 27
10 Computational Details I (K. & Braun, 2007) m = 1 We compute the derivative along the lines of the PLS algorithm 1. w 1 = X y w 1 = X 2. t 1 = Xw 1 t 1 = X w 1 = XX 3. ŷ 1 = P t1 y ŷ 1 = 1 t 1 2 ( t1 y + t 1 yi n ) (In P t1 ) t 1 + P t1 m > 1 We rearrange the algorithm in terms of projections P onto vectors t i. The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 10 / 27
11 Computational Details II (K., Sugiyama, & Braun, 2009) Define empirical scatter matrices s = X y and S = X X w 1,... w m is an orthogonal basis of the Krylov space K m (S, s) = span ( s, Ss,..., S m 1 s ) }{{} =:K m PLS computes orthogonal Gram-Schmidt basis of s, Ss,..., S m 1 s Minimization property: ) 1 β m = arg min y Xβ 2 = K m (KmSK m K m s β K m(s,s) explicit formula for the trace of ŷ m (but not for the derivative itself) The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 11 / 27
12 Summary Computational Details Two equivalent algorithms implemented in the R-package plsdof. my.pls<-pls.model(x,y,compute.dof=true,compute.jacobian) option: runtime scales in p (number of variables) or in n (number of observations). iterative computation of ŷ m projection on Krylov subspaces ŷ m and β m = A? yes no confidence intervals ) yes no for β m? ĉov ( βm = σ 2 AA run time The details can be found in the paper. The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 12 / 27
13 Shape of the DoF-Curve Benchmark data with different correlation structures. training data set description variables examples examples correlation kin (fh) dynamics of a robot arm low boston census data for housing medium cookie near infrared spectra high pls.object<-pls.model(x, y, compute.dof=true) kin (fh) boston cookie Degrees of Freedom DoF(m)=m+1 Degrees of Freedom DoF(m)=m+1 Degrees of Freedom DoF(m)=m number of components number of components number of components The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 13 / 27
14 A Lower Bound for the Degrees of Freedom The lower is the collinearity, the higher are the Degrees of Freedom. Theorem If the largest eigenvalue λ max of the sample correlation matrix S fulfills then 2λ max trace(s) DoF(m = 1) 1 + trace(s) λ max. The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 14 / 27
15 Shape of the DoF-Curve Benchmark data with different correlation structures. training data set description variables examples examples correlation kin (fh) dynamics of a robot arm low boston census data for housing medium cookie near infrared spectra high pls.object<-pls.model(x, y, compute.dof=true) Degrees of Freedom kin (fh) * ******************************* lower bound DoF(m)=m+1 * Degrees of Freedom boston * * * * * * * * * * * * DoF(m)=m+1 lower bound * * Degrees of Freedom *** **** * * * ** cookie * * * * * *** ** ************** DoF(m)=m number of components number of components number of components The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 15 / 27
16 Comparison of Regression Methods: ozone L.A. ozone pollution data p = 12 variables n = 203 observations, n train = 50 training observations Comparison of 1. Partial Least Squares (PLS) 2. Principal Components Regression (PCR) 3. Ridge Regression β ridge = { arg min y Xβ 2 + λ β 2} β The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 16 / 27
17 Comparison of Regression Methods: ozone mean-squared error and model complexity mean squared error PLS PCR RIDGE components PLS PCR degrees of freedom PLS PCR RIDGE There is no difference with respect to mean-squared error. A direct comparison of model parameters (number of components and λ) is not possible. Degrees of Freedom enable a fair model comparison between PLS and PCR. The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 17 / 27
18 Comparison of Regression Methods: ozone training error x o x PLS PCR o x x x o x x x x o x o o o o x o o o o x o x components training error x o x x x x ox x x xo xooo x oo x degrees of freedom,,pls fits closer than PCR. (de Jong) However, there is no clear difference with respect to the Degrees of Freedom. PLS puts focus on more complex models. The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 18 / 27
19 Variable Selection for PLS PLS does not select variables. extensions to sparse PLS 1. thresholding of the weight vectors (Saigo, K.,& Tsuda, 2008) 2. sparsity constraints on the weight vectors (Le Cao et. al., 2008; Chun & Keles, 2010) 3. shrinkage (Kondylis & Whittaker, 2007) Classical approaches 5. bootstrapping R packages pls and ppls 6. hypothesis testing based on the distribution of β The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 19 / 27
20 Approximate distribution of β Recall: Regression coefficients are a non-linear function of y. First order Taylor approximation β β y y }{{} =:A approximate covariance matrix ) ĉov ( β = σ 2 AA The noise level can be estimated via σ 2 = y ŷ 2 n DoF The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 20 / 27
21 ( ( ( ) ( ) ( ) ) ( ) ) ( ) ( ) ( ) ( ( ) ) ( ( ) ((( ( ) ))) ) ( ) ))))) ( ( ( ) ) ) ( ) ( ) ) ) Confidence Intervals for PLS: ozone and tecator implemented in the R-packages plsdof and multcomp cv.object<-pls.cv(x, y, compute.covariance=true) my.multcomp<-glht(cv.object,...) 95% family wise confidence level 1 ( ) 2 (( ))) 3 4 ( 5 ( ) 6 ( ) 7 ( ) 8 ( ) 9 ( ) ( ) ( ) ( ) % family wise confidence level ))))))) )) )))) (((((((( (( ( ((( (( )) ( ) (( ( ( )) ( ) ) ( ) ( ) ( ) ( ) ( ) ( (( ) ))) ( (( ((((((((((((( ((((((( )))))) ( (((( )) ((( )))) The function corrects for multiple comparisons. The computational cost is high for large number of variables. The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 21 / 27
22 Model Selection Comparison of fold cross-validation (gold standard) cv.object<-pls.cv(x, y, k=10) 2. Bayesian Information Criterion with our DoF-estimate bic.object<-pls.ic(x, y, criterion= bic ) 3. Bayesian Information Criterion with the naive estimate DoF=m+1 naive.object<-pls.ic(x, y, criterion= bic, naive=true) Akaike Information Criterion and Minimum Description Length are also available in the R-package. data sets: kin (fh), Boston, cookie The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 22 / 27
23 Prediction Accuracy 10-fold cross-validation Bayesian Information Criterion with our DoF-estimate Bayesian Information Criterion with the naive estimate DoF=m+1 kin (fh) boston 30 cookie test error * test error test error * CV BIC BIC naive CV BIC BIC naive CV BIC BIC naive 1. All three approaches obtain similar accuracy. 2. There is no clear difference between BIC and naive BIC. The plots look similar for the selected Degrees of Freedom. The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 23 / 27
24 Model complexity (selected components) 10-fold cross-validation Bayesian Information Criterion with our DoF-estimate Bayesian Information Criterion with the naive estimate DoF=m+1 kin (fh) boston cookie number of components number of components number of components CV BIC BIC naive 1 CV BIC BIC naive 0 CV BIC BIC naive 1. BIC selects less complex models than naive BIC. 2. There is no clear difference between BIC and CV. The plots look similar for the selected Degrees of Freedom. The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 24 / 27
25 Why not use the naive approach? The naive approach selects more components, but the mean-squared error is not higher. no overfitting? The test error curve can be flat or steep around the optimum x o scaled test error d=50 d=210 o ooo o oo oo o o ooo o o o o oo x ox x x x xxxxxxxx xxxxxxxxxxxxxxxxx ooo oo oo o componenents Depending on the form of the curve, the selection of too complex models does lead to overfitting. More details in the paper. The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 25 / 27
26 Summary Partial Least Squares... typically consumes more than one Degree of Freedom for each component. precise estimate of its intrinsic complexity ( ) ŷm DoF(m) = trace y Its Degrees of Freedom... allow us to compare different regression methods. select less complex models than the naive estimate (when combined with information criteria). Variables can be selected... by constructing approximate confidence intervals. The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 26 / 27
27 References Krämer, N. and Sugiyama, M. (2011). The Degrees of Freedom of Partial Least Squares Regression Journal of the American Statistical Association, in press Krämer, N. and Braun, M. L. (2010). plsdof: Degrees of Freedom and Confidence Intervals for Partial Least Squares R package version Krämer, N., Sugiyama, M. Braun, M.L. (2009). Lanczos Approximations for the Speedup of Kernel Partial Least Squares Regression Twelfth International Conference on Artificial Intelligence and Statistics (AISTATS) Krämer, N., Braun, M.L. (2007) Kernelizing PLS, Degrees of Freedom, and Efficient Model Selection 24th International Conference on Machine Learning (ICML) The Degrees of Freedom of PLS Dr. Nicole Krämer (TUM) 27 / 27
Preface... xi. A Word to the Practitioner... xi The Organization of the Book... xi Required Software... xii Accessing the Supplementary Content...
Contents Preface... xi A Word to the Practitioner... xi The Organization of the Book... xi Required Software... xii Accessing the Supplementary Content... xii Chapter 1 Introducing Partial Least Squares...
More informationGetting Started with Correlated Component Regression (CCR) in XLSTAT-CCR
Tutorial 1 Getting Started with Correlated Component Regression (CCR) in XLSTAT-CCR Dataset for running Correlated Component Regression This tutorial 1 is based on data provided by Michel Tenenhaus and
More informationLecture 2. Review of Linear Regression I Statistics Statistical Methods II. Presented January 9, 2018
Review of Linear Regression I Statistics 211 - Statistical Methods II Presented January 9, 2018 Estimation of The OLS under normality the OLS Dan Gillen Department of Statistics University of California,
More informationRegularized Linear Models in Stacked Generalization
Regularized Linear Models in Stacked Generalization Sam Reid and Greg Grudic Department of Computer Science University of Colorado at Boulder USA June 11, 2009 Reid & Grudic (Univ. of Colo. at Boulder)
More informationProfessor Dr. Gholamreza Nakhaeizadeh. Professor Dr. Gholamreza Nakhaeizadeh
Statistic Methods in in Data Mining Business Understanding Data Understanding Data Preparation Deployment Modelling Evaluation Data Mining Process (Part 2) 2) Professor Dr. Gholamreza Nakhaeizadeh Professor
More informationInvestigation in to the Application of PLS in MPC Schemes
Ian David Lockhart Bogle and Michael Fairweather (Editors), Proceedings of the 22nd European Symposium on Computer Aided Process Engineering, 17-20 June 2012, London. 2012 Elsevier B.V. All rights reserved
More informationImproving CERs building
Improving CERs building Getting Rid of the R² tyranny Pierre Foussier pmf@3f fr.com ISPA. San Diego. June 2010 1 Why abandon the OLS? The ordinary least squares (OLS) aims to build a CER by minimizing
More informationPublished: 14 October 2014
Electronic Journal of Applied Statistical Analysis EJASA, Electron. J. App. Stat. Anal. http://siba-ese.unisalento.it/index.php/ejasa/index e-issn: 2070-5948 DOI: 10.1285/i20705948v7n2p343 A note on ridge
More informationPARTIAL LEAST SQUARES: APPLICATION IN CLASSIFICATION AND MULTIVARIABLE PROCESS DYNAMICS IDENTIFICATION
PARIAL LEAS SQUARES: APPLICAION IN CLASSIFICAION AND MULIVARIABLE PROCESS DYNAMICS IDENIFICAION Seshu K. Damarla Department of Chemical Engineering National Institute of echnology, Rourkela, India E-mail:
More informationTopic 5 Lecture 3 Estimating Policy Effects via the Simple Linear. Regression Model (SLRM) and the Ordinary Least Squares (OLS) Method
Econometrics for Health Policy, Health Economics, and Outcomes Research Topic 5 Lecture 3 Estimating Policy Effects via the Simple Linear Regression Model (SLRM) and the Ordinary Least Squares (OLS) Method
More information5. CONSTRUCTION OF THE WEIGHT-FOR-LENGTH AND WEIGHT-FOR- HEIGHT STANDARDS
5. CONSTRUCTION OF THE WEIGHT-FOR-LENGTH AND WEIGHT-FOR- HEIGHT STANDARDS 5.1 Indicator-specific methodology The construction of the weight-for-length (45 to 110 cm) and weight-for-height (65 to 120 cm)
More informationIntegrating remote sensing and ground monitoring data to improve estimation of PM 2.5 concentrations for chronic health studies
Integrating remote sensing and ground monitoring data to improve estimation of PM 2.5 concentrations for chronic health studies Chris Paciorek and Yang Liu Departments of Biostatistics and Environmental
More informationPLS score-loading correspondence and a bi-orthogonal factorization
PLS score-loading correspondence and a bi-orthogonal factorization Rolf Ergon elemark University College P.O.Box, N-9 Porsgrunn, Norway e-mail: rolf.ergon@hit.no telephone: ++ 7 7 telefax: ++ 7 7 Published
More informationPARTIAL LEAST SQUARES: WHEN ORDINARY LEAST SQUARES REGRESSION JUST WON T WORK
PARTIAL LEAST SQUARES: WHEN ORDINARY LEAST SQUARES REGRESSION JUST WON T WORK Peter Bartell JMP Systems Engineer peter.bartell@jmp.com WHEN OLS JUST WON T WORK? OLS (Ordinary Least Squares) in JMP/JMP
More informationAn Introduction to Partial Least Squares Regression
An Introduction to Partial Least Squares Regression Randall D. Tobias, SAS Institute Inc., Cary, NC Abstract Partial least squares is a popular method for soft modelling in industrial applications. This
More informationCollective Traffic Prediction with Partially Observed Traffic History using Location-Based Social Media
Collective Traffic Prediction with Partially Observed Traffic History using Location-Based Social Media Xinyue Liu, Xiangnan Kong, Yanhua Li Worcester Polytechnic Institute February 22, 2017 1 / 34 About
More informationStatistical Learning Examples
Statistical Learning Examples Genevera I. Allen Statistics 640: Statistical Learning August 26, 2013 (Stat 640) Lecture 1 August 26, 2013 1 / 19 Example: Microarrays arrays High-dimensional: Goals: Measures
More informationAnalysis of Big Data Streams to Obtain Braking Reliability Information July 2013, for 2017 Train Protection 1 / 25
Analysis of Big Data Streams to Obtain Braking Reliability Information for Train Protection Systems Prof. Dr. Raphael Pfaff Aachen University of Applied Sciences pfaff@fh-aachen.de www.raphaelpfaff.net
More informationWorkshop on Frame Theory and Sparse Representation for Complex Data June 1, 2017
Workshop on Frame Theory and Sparse Representation for Complex Data June 1, 2017 Xiaoming Huo Georgia Institute of Technology School of industrial and systems engineering I. Statistical Dependence II.
More informationDomain-invariant Partial Least Squares (di-pls) Regression: A novel method for unsupervised and semi-supervised calibration model adaptation
Domain-invariant Partial Least Squares (di-pls) Regression: A novel method for unsupervised and semi-supervised calibration model adaptation R. Nikzad-Langerodi W. Zellinger E. Lughofer T. Reischer 2 S.
More information9.2 User s Guide SAS/STAT. The PLS Procedure. (Book Excerpt) SAS Documentation
SAS/STAT 9.2 User s Guide The PLS Procedure (Book Excerpt) SAS Documentation This document is an individual chapter from SAS/STAT 9.2 User s Guide. The correct bibliographic citation for the complete manual
More informationArticle: Sulfur Testing VPS Quality Approach By Dr Sunil Kumar Laboratory Manager Fujairah, UAE
Article: Sulfur Testing VPS Quality Approach By Dr Sunil Kumar Laboratory Manager Fujairah, UAE 26th September 2017 For over a decade, both regional ECA and global sulphur limits within marine fuels have
More informationKINEMATICAL SUSPENSION OPTIMIZATION USING DESIGN OF EXPERIMENT METHOD
Jurnal Mekanikal June 2014, No 37, 16-25 KINEMATICAL SUSPENSION OPTIMIZATION USING DESIGN OF EXPERIMENT METHOD Mohd Awaluddin A Rahman and Afandi Dzakaria Faculty of Mechanical Engineering, Universiti
More informationSharif University of Technology. Graduate School of Management and Economics. Econometrics I. Fall Seyed Mahdi Barakchian
Sharif University of Technology Graduate School of Management and Economics Econometrics I Fall 2010 Seyed Mahdi Barakchian Textbook: Wooldridge, J., Introductory Econometrics: A Modern Approach, South
More informationOptimal Vehicle to Grid Regulation Service Scheduling
Optimal to Grid Regulation Service Scheduling Christian Osorio Introduction With the growing popularity and market share of electric vehicles comes several opportunities for electric power utilities, vehicle
More informationBayesian Trajectory Optimization for Magnetic Resonance Imaging Sequences
Bayesian Trajectory Optimization for Magnetic Resonance Imaging Sequences Matthias Seeger Saarland University and MPI for Informatics, Saarbrücken Joint work with Hannes Nickisch, Rolf Pohmann, Bernhard
More informationOptimization Methodology for CVT Ratio Scheduling with Consideration of Both Engine and CVT Efficiency
Western Michigan University ScholarWorks at WMU Master's Theses Graduate College 12-2016 Optimization Methodology for CVT Ratio Scheduling with Consideration of Both Engine and CVT Efficiency Steven Beuerle
More informationCalibration. DOE & Statistical Modeling
ETAS Webinar - ASCMO Calibration. DOE & Statistical Modeling Injection Consumption Ignition Torque AFR HC EGR P-rail NOx Inlet-cam Outlet-cam 1 1 Soot T-exhaust Roughness What is Design of Experiments?
More informationPredicting Solutions to the Optimal Power Flow Problem
Thomas Navidi Suvrat Bhooshan Aditya Garg Abstract Predicting Solutions to the Optimal Power Flow Problem This paper discusses an implementation of gradient boosting regression to predict the output of
More informationTechnical Papers supporting SAP 2009
Technical Papers supporting SAP 29 A meta-analysis of boiler test efficiencies to compare independent and manufacturers results Reference no. STP9/B5 Date last amended 25 March 29 Date originated 6 October
More informationSAS/STAT 13.1 User s Guide. The PLS Procedure
SAS/STAT 13.1 User s Guide The PLS Procedure This document is an individual chapter from SAS/STAT 13.1 User s Guide. The correct bibliographic citation for the complete manual is as follows: SAS Institute
More informationChapter 5 ESTIMATION OF MAINTENANCE COST PER HOUR USING AGE REPLACEMENT COST MODEL
Chapter 5 ESTIMATION OF MAINTENANCE COST PER HOUR USING AGE REPLACEMENT COST MODEL 87 ESTIMATION OF MAINTENANCE COST PER HOUR USING AGE REPLACEMENT COST MODEL 5.1 INTRODUCTION Maintenance is usually carried
More informationAntonio Olmos Priyalatha Govindasamy Research Methods & Statistics University of Denver
Antonio Olmos Priyalatha Govindasamy Research Methods & Statistics University of Denver American Evaluation Association Conference, Chicago, Ill, November 2015 AEA 2015, Chicago Ill 1 Paper overview Propensity
More informationDOT HS Summary of Statistical Findings November Statistical Methodology to Make Early Estimates of Motor Vehicle Traffic Fatalities
TRAFFIC SAFETY FACTS Research Note DOT HS 811 123 Summary of Statistical Findings November 2010 Statistical Methodology to Make Early Estimates of Motor Vehicle Traffic Fatalities Highlights Beginning
More informationPREDICTION OF FUEL CONSUMPTION
PREDICTION OF FUEL CONSUMPTION OF AGRICULTURAL TRACTORS S. C. Kim, K. U. Kim, D. C. Kim ABSTRACT. A mathematical model was developed to predict fuel consumption of agricultural tractors using their official
More informationRobust alternatives to best linear unbiased prediction of complex traits
Robust alternatives to best linear unbiased prediction of complex traits WHY BEST LINEAR UNBIASED PREDICTION EASY TO EXPLAIN FLEXIBLE AMENDABLE WELL UNDERSTOOD FEASIBLE UNPRETENTIOUS NORMALITY IS IMPLICIT
More informationPUBLICATIONS Silvia Ferrari February 24, 2017
PUBLICATIONS Silvia Ferrari February 24, 2017 [1] Cordeiro, G.M., Ferrari, S.L.P. (1991). A modified score test statistic having chi-squared distribution to order n 1. Biometrika, 78, 573-582. [2] Cordeiro,
More informationForecasting China s Inflation in a Data-Rich. Environment
Forecasting China s Inflation in a Data-Rich Environment Ching-Yi Lin Department of Economics, National Tsing Hua University Chun Wang Department of Economics, Brooklyn College, CUNY Abstract Inflation
More informationForecast the charging power demand for an electric vehicle. Dr. Wilson Maluenda, FH Vorarlberg; Philipp Österle, Illwerke VKW;
Forecast the charging power demand for an electric vehicle Dr. Wilson Maluenda, FH Vorarlberg; Philipp Österle, Illwerke VKW; Vienna, Bregenz; Austria 11.03.2015 Content Abstract... 1 Motivation... 2 Challenges...
More informationOnline Learning and Optimization for Smart Power Grid
1 2016 IEEE PES General Meeting Panel on Domain-Specific Big Data Analytics Tools in Power Systems Online Learning and Optimization for Smart Power Grid Seung-Jun Kim Department of Computer Sci. and Electrical
More informationRegression Models Course Project, 2016
Regression Models Course Project, 2016 Venkat Batchu July 13, 2016 Executive Summary In this report, mtcars data set is explored/analyzed for relationship between outcome variable mpg (miles for gallon)
More informationCost-Efficiency by Arash Method in DEA
Applied Mathematical Sciences, Vol. 6, 2012, no. 104, 5179-5184 Cost-Efficiency by Arash Method in DEA Dariush Khezrimotlagh*, Zahra Mohsenpour and Shaharuddin Salleh Department of Mathematics, Faculty
More informationOnline Learning and Optimization for Smart Power Grid
1 2016 IEEE PES General Meeting Panel on Domain-Specific Big Data Analytics Tools in Power Systems Online Learning and Optimization for Smart Power Grid Seung-Jun Kim Department of Computer Sci. and Electrical
More informationUnderstanding the benefits of using a digital valve controller. Mark Buzzell Business Manager, Metso Flow Control
Understanding the benefits of using a digital valve controller Mark Buzzell Business Manager, Metso Flow Control Evolution of Valve Positioners Digital (Next Generation) Digital (First Generation) Analog
More informationLondon calling (probably)
London calling (probably) Parameters and stochastic behaviour of braking force generation and transmission Prof. Dr. Raphael Pfaff Aachen University of Applied Sciences pfaff@fh-aachen.de www.raphaelpfaff.net
More informationIntroduction. Materials and Methods. How to Estimate Injection Percentage
How to Estimate Injection Percentage Introduction The Marel IN33-3 injector for pork bellies is a 5 needle, low-pressure conveyor type machine which utilizes a 3-gpm positive displacement pump and control
More informationSvante Wold, Research Group for Chemometrics, Institute of Chemistry, Umeå University, S Umeå, Sweden
Submitted version, June 2004 The PLS method -- partial least squares projections to latent structures -- and its applications in industrial RDP (research, development, and production). Svante Wold, Research
More informationACCIDENT MODIFICATION FACTORS FOR MEDIAN WIDTH
APPENDIX G ACCIDENT MODIFICATION FACTORS FOR MEDIAN WIDTH INTRODUCTION Studies on the effect of median width have shown that increasing width reduces crossmedian crashes, but the amount of reduction varies
More informationLinking the Indiana ISTEP+ Assessments to the NWEA MAP Growth Tests. February 2017 Updated November 2017
Linking the Indiana ISTEP+ Assessments to the NWEA MAP Growth Tests February 2017 Updated November 2017 2017 NWEA. All rights reserved. No part of this document may be modified or further distributed without
More informationProbabilistic Modeling of Fatigue Damage in Steel Box-Girder Bridges Subject to Stochastic Vehicle Loads
505 Probabilistic Modeling of Fatigue Damage in Steel Box-Girder Bridges Subject to Stochastic Vehicle Loads Yuan Luo 1), Dong-huang Yan 1) and Nai-wei Lu 2) 1) School of Civil Engineering and Architecture,
More informationEnhanced gear efficiency calculation including contact analysis results and drive cycle consideration
Enhanced gear efficiency calculation including contact analysis results and drive cycle consideration Dipl.-Ing. J. Langhart, KISSsoft AG, CH-Bubikon; M. Sc. T. Panero, KISSsoft AG, CH-Bubikon Abstract
More informationLinking the Alaska AMP Assessments to NWEA MAP Tests
Linking the Alaska AMP Assessments to NWEA MAP Tests February 2016 Introduction Northwest Evaluation Association (NWEA ) is committed to providing partners with useful tools to help make inferences from
More informationDynamic Modeling of Large Complex Hydraulic System Based on Virtual Prototyping Gui-bo YU, Jian-zhuang ZHI *, Li-jun CAO and Qiao MA
2018 International Conference on Computer, Electronic Information and Communications (CEIC 2018) ISBN: 978-1-60595-557-5 Dynamic Modeling of Large Complex Hydraulic System Based on Virtual Prototyping
More informationME scope Application Note 29 FEA Model Updating of an Aluminum Plate
ME scope Application Note 29 FEA Model Updating of an Aluminum Plate NOTE: You must have a package with the VES-4500 Multi-Reference Modal Analysis and VES-8000 FEA Model Updating options enabled to reproduce
More informationStopping criteria in iterative methods a miscellaneous issue?
Stopping criteria in iterative methods a miscellaneous issue? Zdeněk Strakoš and Chris C. Paige Institute of Computer Science AS CR, McGill University, Montreal, Canada. GAMM Annual Conference, Berlin,
More informationRobust and Classical PLS Regression Compared
Robust and Classical PLS Regression Compared Bettina Liebmann a *, Peter Filzmoser b and Kurt Varmuza a ------------------------------- * Correspondence to: B. Liebmann, Laboratory for Chemometrics, Institute
More informationAPPLICATION OF A PCA MODEL APPROACH FOR MISFIRE MONITORING. Paul J. King 1 and Keith J. Burnham 2
APPLICATION OF A PCA MODEL APPROACH FOR MISFIRE MONITORING Paul J. King 1 and Keith J. Burnham 2 1 Powertrain Control Systems and Calibration, Jaguar Cars Limited, Coventry, CV3 4BJ, U.K. 2 Control Theory
More informationNRC Non-Destructive Examination Research
NRC Non-Destructive Examination Research Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Jeff Hixon October 22, 2009 Why NDE? U.S. Nuclear Regulatory Commission 2 Why NDE? U.S.
More informationUSE OF PLS COMPONENTS TO IMPROVE CLASSIFICATION ON BUSINESS DECISION MAKING
USE OF PLS COMPONENTS TO IMPROVE CLASSIFICATION ON BUSINESS DECISION MAKING José C. Vega Vilca, Aniel Nieves-González and Roxana Aparicio Institute of Statistics and Computer Information Systems, University
More informationLinking the Virginia SOL Assessments to NWEA MAP Growth Tests *
Linking the Virginia SOL Assessments to NWEA MAP Growth Tests * *As of June 2017 Measures of Academic Progress (MAP ) is known as MAP Growth. March 2016 Introduction Northwest Evaluation Association (NWEA
More informationLinking the Georgia Milestones Assessments to NWEA MAP Growth Tests *
Linking the Georgia Milestones Assessments to NWEA MAP Growth Tests * *As of June 2017 Measures of Academic Progress (MAP ) is known as MAP Growth. February 2016 Introduction Northwest Evaluation Association
More informationHASIL OUTPUT SPSS. Reliability Scale: ALL VARIABLES
139 HASIL OUTPUT SPSS Reliability Scale: ALL VARIABLES Case Processing Summary N % 100 100.0 Cases Excluded a 0.0 Total 100 100.0 a. Listwise deletion based on all variables in the procedure. Reliability
More informationEmbedded Torque Estimator for Diesel Engine Control Application
2004-xx-xxxx Embedded Torque Estimator for Diesel Engine Control Application Peter J. Maloney The MathWorks, Inc. Copyright 2004 SAE International ABSTRACT To improve vehicle driveability in diesel powertrain
More informationLinking the North Carolina EOG Assessments to NWEA MAP Growth Tests *
Linking the North Carolina EOG Assessments to NWEA MAP Growth Tests * *As of June 2017 Measures of Academic Progress (MAP ) is known as MAP Growth. March 2016 Introduction Northwest Evaluation Association
More informationFuel Economy and Safety
Fuel Economy and Safety A Reexamination under the U.S. Footprint-Based Fuel Economy Standards Jiaxi Wang University of California, Irvine Abstract The purpose of this study is to reexamine the tradeoff
More informationBayes Factors. Structural Equation Models (SEMs): Schwarz BIC and Other Approximations
Bayes Factors in Structural Equation Models (SEMs): Schwarz BIC and Other Approximations Kenneth A. Bollen University of North Carolina, Chapel Hill Surajit Ray SAMSI and University of North Carolina,
More informationUse of Flow Network Modeling for the Design of an Intricate Cooling Manifold
Use of Flow Network Modeling for the Design of an Intricate Cooling Manifold Neeta Verma Teradyne, Inc. 880 Fox Lane San Jose, CA 94086 neeta.verma@teradyne.com ABSTRACT The automatic test equipment designed
More informationLinking the Kansas KAP Assessments to NWEA MAP Growth Tests *
Linking the Kansas KAP Assessments to NWEA MAP Growth Tests * *As of June 2017 Measures of Academic Progress (MAP ) is known as MAP Growth. February 2016 Introduction Northwest Evaluation Association (NWEA
More informationWavelet-PLS Regression: Application to Oil Production Data
Wavelet-PLS Regression: Application to Oil Production Data Benammou Saloua 1, Kacem Zied 1, Kortas Hedi 1, and Dhifaoui Zouhaier 1 1 Computational Mathematical Laboratory, saloua.benammou@yahoo.fr 2 ZiedKacem2004@yahoo.fr
More informationData envelopment analysis with missing values: an approach using neural network
IJCSNS International Journal of Computer Science and Network Security, VOL.17 No.2, February 2017 29 Data envelopment analysis with missing values: an approach using neural network B. Dalvand, F. Hosseinzadeh
More informationLinking the Mississippi Assessment Program to NWEA MAP Tests
Linking the Mississippi Assessment Program to NWEA MAP Tests February 2017 Introduction Northwest Evaluation Association (NWEA ) is committed to providing partners with useful tools to help make inferences
More informationAccelerating the Development of Expandable Liner Hanger Systems using Abaqus
Accelerating the Development of Expandable Liner Hanger Systems using Abaqus Ganesh Nanaware, Tony Foster, Leo Gomez Baker Hughes Incorporated Abstract: Developing an expandable liner hanger system for
More informationLinking the New York State NYSTP Assessments to NWEA MAP Growth Tests *
Linking the New York State NYSTP Assessments to NWEA MAP Growth Tests * *As of June 2017 Measures of Academic Progress (MAP ) is known as MAP Growth. March 2016 Introduction Northwest Evaluation Association
More informationSTUDY OF AIRBAG EFFECTIVENESS IN HIGH SEVERITY FRONTAL CRASHES
STUDY OF AIRBAG EFFECTIVENESS IN HIGH SEVERITY FRONTAL CRASHES Jeya Padmanaban (JP Research, Inc., Mountain View, CA, USA) Vitaly Eyges (JP Research, Inc., Mountain View, CA, USA) ABSTRACT The primary
More informationAnalysis of Partial Least Squares for Pose-Invariant Face Recognition
Analysis of Partial Least Squares for Pose-Invariant Face Recognition Mika Fischer Hazım Kemal Ekenel, Rainer Stiefelhagen mika.fischer@kit.edu ekenel@{kit.edu,itu.edu.tr} rainer.stiefelhagen@kit.edu Karlsruhe
More informationLinking the Florida Standards Assessments (FSA) to NWEA MAP
Linking the Florida Standards Assessments (FSA) to NWEA MAP October 2016 Introduction Northwest Evaluation Association (NWEA ) is committed to providing partners with useful tools to help make inferences
More informationSupporting information
Supporting information Table S1. Tensor parameters and shift of metal ion from starting position for the 10 subsets (from a total of 100) resulting in the largest number of peak pairings, using 2D and
More informationLinking the Indiana ISTEP+ Assessments to NWEA MAP Tests
Linking the Indiana ISTEP+ Assessments to NWEA MAP Tests February 2017 Introduction Northwest Evaluation Association (NWEA ) is committed to providing partners with useful tools to help make inferences
More informationA Battery Smart Sensor and Its SOC Estimation Function for Assembled Lithium-Ion Batteries
R1-6 SASIMI 2015 Proceedings A Battery Smart Sensor and Its SOC Estimation Function for Assembled Lithium-Ion Batteries Naoki Kawarabayashi, Lei Lin, Ryu Ishizaki and Masahiro Fukui Graduate School of
More informationRelating your PIRA and PUMA test marks to the national standard
Relating your PIRA and PUMA test marks to the national standard We have carried out a detailed statistical analysis between the results from the PIRA and PUMA tests for Year 2 and Year 6 and the scaled
More informationRelating your PIRA and PUMA test marks to the national standard
Relating your PIRA and PUMA test marks to the national standard We have carried out a detailed statistical analysis between the results from the PIRA and PUMA tests for Year 2 and Year 6 and the scaled
More informationStatistics and Quantitative Analysis U4320. Segment 8 Prof. Sharyn O Halloran
Statistics and Quantitative Analysis U4320 Segment 8 Prof. Sharyn O Halloran I. Introduction A. Overview 1. Ways to describe, summarize and display data. 2.Summary statements: Mean Standard deviation Variance
More informationINVITED REVIEW PAPER. Faisal Ahmed*, Lae-Hyun Kim**, and Yeong-Koo Yeo*,
Korean J. Chem. Eng., 30(1), 11-19 (2013) DOI: 10.1007/s11814-012-0107-z INVITED REVIEW PAPER Statistical data modeling based on partial least squares: Application to melt index predictions in high density
More informationSmart Operation for AC Distribution Infrastructure Involving Hybrid Renewable Energy Sources
Milano (Italy) August 28 - September 2, 211 Smart Operation for AC Distribution Infrastructure Involving Hybrid Renewable Energy Sources Ahmed A Mohamed, Mohamed A Elshaer and Osama A Mohammed Energy Systems
More informationPREDICTION OF REMAINING USEFUL LIFE OF AN END MILL CUTTER SEOW XIANG YUAN
PREDICTION OF REMAINING USEFUL LIFE OF AN END MILL CUTTER SEOW XIANG YUAN Report submitted in partial fulfillment of the requirements for the award of the degree of Bachelor of Engineering (Hons.) in Manufacturing
More informationInvestigating the Concordance Relationship Between the HSA Cut Scores and the PARCC Cut Scores Using the 2016 PARCC Test Data
Investigating the Concordance Relationship Between the HSA Cut Scores and the PARCC Cut Scores Using the 2016 PARCC Test Data A Research Report Submitted to the Maryland State Department of Education (MSDE)
More information2018 Linking Study: Predicting Performance on the Performance Evaluation for Alaska s Schools (PEAKS) based on MAP Growth Scores
2018 Linking Study: Predicting Performance on the Performance Evaluation for Alaska s Schools (PEAKS) based on MAP Growth Scores June 2018 NWEA Psychometric Solutions 2018 NWEA. MAP Growth is a registered
More informationTABLE 4.1 POPULATION OF 100 VALUES 2
TABLE 4. POPULATION OF 00 VALUES WITH µ = 6. AND = 7.5 8. 6.4 0. 9.9 9.8 6.6 6. 5.7 5. 6.3 6.7 30.6.6.3 30.0 6.5 8. 5.6 0.3 35.5.9 30.7 3.. 9. 6. 6.8 5.3 4.3 4.4 9.0 5.0 9.9 5. 0.8 9.0.9 5.4 7.3 3.4 38..6
More informationSPEED AND TORQUE CONTROL OF AN INDUCTION MOTOR WITH ANN BASED DTC
SPEED AND TORQUE CONTROL OF AN INDUCTION MOTOR WITH ANN BASED DTC Fatih Korkmaz Department of Electric-Electronic Engineering, Çankırı Karatekin University, Uluyazı Kampüsü, Çankırı, Turkey ABSTRACT Due
More informationIDEA for GOES-R ABI. Presented by S. Kondragunta, NESDIS/STAR. Team Members: R. Hoff and H. Zhang, UMBC
IDEA for GOES-R ABI Presented by S. Kondragunta, NESDIS/STAR Team Members: R. Hoff and H. Zhang, UMBC 1 Project Summary Use operational MODIS, GOES Aerosol Optical Depth (AOD) products, and OMI/GOME-2
More informationA UNIFYING VIEW ON MULTI-STEP FORECASTING USING AN AUTOREGRESSION
doi: 10.1111/j.1467-6419.2009.00581.x A UNIFYING VIEW ON MULTI-STEP FORECASTING USING AN AUTOREGRESSION Philip Hans Franses and Rianne Legerstee Econometric Institute and Tinbergen Institute, Erasmus University
More informationAssignment 3 solutions
Assignment 3 solutions Question 1: SVM on the OJ data (a) [2 points] Create a training set containing a random sample of 800 observations, and a test set containing the remaining observations. library(islr)
More informationMORSE: MOdel-based Real-time Systems Engineering. Reducing physical testing in the calibration of diagnostic and driveabilty features
MORSE: MOdel-based Real-time Systems Engineering Reducing physical testing in the calibration of diagnostic and driveabilty features Mike Dempsey Claytex Future Powertrain Conference 2017 MORSE project
More informationFast and Robust Optimization Approaches for Pedestrian Detection
Fast and Robust Optimization Approaches for Pedestrian Detection Victor Hugo Cunha de Melo, David Menotti (Co-advisor), William Robson Schwartz (Advisor) Computer Science Department, Universidade Federal
More informationStatistical Applications in Genetics and Molecular Biology
Statistical Applications in Genetics and Molecular Biology Volume 3, Issue 1 2004 Article 33 PLS Dimension Reduction for Classification with Microarray Data Anne-Laure Boulesteix Department of Statistics,
More informationEstimation of Unmeasured DOF s on a Scaled Model of a Blade Structure
Estimation of Unmeasured DOF s on a Scaled Model of a Blade Structure Anders Skafte 1, Rune Brincker 2 ABSTRACT This paper presents a new expansion technique which enables to predict mode shape coordinates
More informationExperimental analysis of a contact patch form of a rolling tire: influence of speed, wheel load, camber and slip angle
Experimental analysis of a contact patch form of a rolling tire: influence of speed, wheel load, camber and slip angle Dipl.-Ing. Pavel Sarkisov Prof. Dr.-Ing. Günther Prokop Dipl.-Ing. Steffen Drossel
More informationIndex. Cambridge University Press Applied Nonparametric Econometrics Daniel J. Henderson and Christopher F. Parmeter.
additive nonparametric models, 254 265 additively separable regression model, 336 337 Afriat conditions, 333 age-earnings regression, 5 6 Ahmad, I.A., 101, 107 Aitchison, J., 189 190, 198 199, 206 Aitken,
More informationA Distributed Neurocomputing Approach for Infrasound Event Classification
A Distributed Neurocomputing Approach for Infrasound Event Classification Fredric M. Ham, Ph.D., FIEEE Harris Professor of Electrical Engineering Director of the Information Processing Laboratory Florida
More informationRoad Surface characteristics and traffic accident rates on New Zealand s state highway network
Road Surface characteristics and traffic accident rates on New Zealand s state highway network Robert Davies Statistics Research Associates http://www.statsresearch.co.nz Joint work with Marian Loader,
More information