Non Linear Regression Y i = f(  x i ) +  i Marco Lattuada Swiss Federal Institute of Technology - ETH Institut für Chemie und Bioingenieurwissenschaften.

Slides:



Advertisements
Similar presentations
Programming Tips: While Loops and Comparisons 1Daniel Baur / Numerical Methods for Chemical Engineerse Daniel Baur ETH Zurich, Institut für Chemie- und.
Advertisements

Boundary Value Problems and Partial Differential Equations (PDEs)
11-1 Empirical Models Many problems in engineering and science involve exploring the relationships between two or more variables. Regression analysis.
Fixed point iterations and solution of non-linear functions
Partial Differential Equations (PDEs) 1Daniel Baur / Numerical Methods for Chemical Engineers / BVP and PDE Daniel Baur ETH Zurich, Institut für Chemie-
Numerical Differentiation and Quadrature (Integration) 1Daniel Baur / Numerical Methods for Chemical Engineers / Numerical Quadrature Daniel Baur ETH Zurich,
Linear Regression 1Daniel Baur / Numerical Methods for Chemical Engineers / Linear Regression Daniel Baur ETH Zurich, Institut für Chemie- und Bioingenieurwissenschaften.
12-1 Multiple Linear Regression Models Introduction Many applications of regression analysis involve situations in which there are more than.
12 Multiple Linear Regression CHAPTER OUTLINE
Nonlinear Regression Ecole Nationale Vétérinaire de Toulouse Didier Concordet ECVPT Workshop April 2011 Can be downloaded at
Systems of Linear Equations
Response Surface Method Principle Component Analysis
Linear Systems of Equations Ax = b Marco Lattuada Swiss Federal Institute of Technology - ETH Institut für Chemie und Bioingenieurwissenschaften ETH Hönggerberg/
Jump to first page STATISTICAL INFERENCE Statistical Inference uses sample data and statistical procedures to: n Estimate population parameters; or n Test.
/k 2DS00 Statistics 1 for Chemical Engineering lecture 5.
Response Surfaces max(S(  )) Marco Lattuada Swiss Federal Institute of Technology - ETH Institut für Chemie und Bioingenieurwissenschaften ETH Hönggerberg/
Non Isothermal Chemical Reactor ChE 479 Alexander Couzis.
Engineering Computation Curve Fitting 1 Curve Fitting By Least-Squares Regression and Spline Interpolation Part 7.
Ordinary Differential Equations (ODEs) 1Daniel Baur / Numerical Methods for Chemical Engineers / Implicit ODE Solvers Daniel Baur ETH Zurich, Institut.
Linear Regression Y i =  0 +  1 x i +  i Marco Lattuada Swiss Federal Institute of Technology - ETH Institut für Chemie und Bioingenieurwissenschaften.
Ordinary Differential Equations (ODEs)
Ordinary Differential Equations (ODEs) 1Daniel Baur / Numerical Methods for Chemical Engineers / Explicit ODE Solvers Daniel Baur ETH Zurich, Institut.
Ordinary Differential Equations (ODEs) 1Daniel Baur / Numerical Methods for Chemical Engineers / Implicit ODE Solvers Daniel Baur ETH Zurich, Institut.
Ordinary Differential Equations (ODEs) 1Daniel Baur / Numerical Methods for Chemical Engineers / Explicit ODE Solvers Daniel Baur ETH Zurich, Institut.
Principles of the Global Positioning System Lecture 10 Prof. Thomas Herring Room A;
Correlation & Regression
Colorado Center for Astrodynamics Research The University of Colorado STATISTICAL ORBIT DETERMINATION Project Report Unscented kalman Filter Information.
Least-Squares Regression
Introduction to Linear Regression and Correlation Analysis
Systems of Linear Equations Iterative Methods
Process modelling and optimization aid FONTEIX Christian Professor of Chemical Engineering Polytechnical National Institute of Lorraine Chemical Engineering.
Chapter 14 – Correlation and Simple Regression Math 22 Introductory Statistics.
Stats for Engineers Lecture 9. Summary From Last Time Confidence Intervals for the mean t-tables Q Student t-distribution.
Analytical vs. Numerical Minimization Each experimental data point, l, has an error, ε l, associated with it ‣ Difference between the experimentally measured.
1 Chapter 3 Multiple Linear Regression Multiple Regression Models Suppose that the yield in pounds of conversion in a chemical process depends.
Boundary Value Problems and Least Squares Minimization
V. Nonlinear Regression Objective-Function Surfaces Thus far, we have: Parameterized the forward model Obtained head and flow observations and their weights.
2014. Engineers often: Regress data  Analysis  Fit to theory  Data reduction Use the regression of others  Antoine Equation  DIPPR We need to be.
Chapter 14 Inference for Regression AP Statistics 14.1 – Inference about the Model 14.2 – Predictions and Conditions.
AP Statistics Chapter 15 Notes. Inference for a Regression Line Goal: To determine if there is a relationship between two quantitative variables. Goal:
Inference for Regression Simple Linear Regression IPS Chapter 10.1 © 2009 W.H. Freeman and Company.
AP Statistics Chapter 15 Notes. Inference for a Regression Line Goal: To determine if there is a relationship between two quantitative variables. –i.e.
Bashkir State Univerity The Chair of Mathematical Modeling , Ufa, Zaki Validi str. 32 Phone: ,
1 Multiple Regression A single numerical response variable, Y. Multiple numerical explanatory variables, X 1, X 2,…, X k.
Identification and Estimation of the Influential Parameters in Bioreaction Systems Mordechai Shacham Ben Gurion University of the Negev Beer-Sheva, Israel.
Quadrature rules 1Michael Sokolov / Numerical Methods for Chemical Engineers / Numerical Quadrature Michael Sokolov ETH Zurich, Institut für Chemie- und.
Numerical Differentiation and Quadrature (Integration)
Solution of Nonlinear Functions
IX. Transient Model Nonlinear Regression and Statistical Analysis.
AP STATISTICS LESSON 14 – 1 ( DAY 1 ) INFERENCE ABOUT THE MODEL.
Ordinary Differential Equations (ODEs) 1Michael Sokolov / Numerical Methods for Chemical Engineers / Explicit ODE Solvers Michael Sokolov ETH Zurich, Institut.
Principal Component Analysis (PCA)
Linear Regression 1Michael Sokolov / Numerical Methods for Chemical Engineers / Linear Regression Michael Sokolov ETH Zurich, Institut für Chemie- und.
Linear Systems of Equations Iterative and Relaxation Methods Ax = b Marco Lattuada Swiss Federal Institute of Technology - ETH Institut für Chemie und.
Nonlinear Regression 1Michael Sokolov / Numerical Methods for Chemical Engineers / Nonlinear Regression Michael Sokolov ETH Zurich, Institut für Chemie-
Regression Analysis AGEC 784.
Lecture 11 Chemical Reaction Engineering (CRE) is the field that studies the rates and mechanisms of chemical reactions and the design of the reactors.
Zero of a Nonlinear Function f(x) = 0
Lecture 11 Chemical Reaction Engineering (CRE) is the field that studies the rates and mechanisms of chemical reactions and the design of the reactors.
Statistical Methods For Engineers
Linear Regression.
Zero of a Nonlinear System of Algebraic Equations f(x) = 0
Linear regression Fitting a straight line to observations.
Nonlinear regression.
6.7 Practical Problems with Curve Fitting simple conceptual problems
6.1 Introduction to Chi-Square Space
Regression Statistics
CHAPTER Five: Collection & Analysis of Rate Data
Chapter 14 Inference for Regression
Multiple linear regression
Presentation transcript:

Non Linear Regression Y i = f(  x i ) +  i Marco Lattuada Swiss Federal Institute of Technology - ETH Institut für Chemie und Bioingenieurwissenschaften ETH Hönggerberg/ HCI F135 – Zürich (Switzerland) Marco Lattuada Swiss Federal Institute of Technology - ETH Institut für Chemie und Bioingenieurwissenschaften ETH Hönggerberg/ HCI F135 – Zürich (Switzerland)

Marco Lattuada – Statistical and Numerical Methods for Chemical Engineers Nonlinear Regressions – Page # 2 Puromycin Description: Puromycin is an antibiotic used by scientists in bio-research to select cells modified by genetic engineering. Mechanism of action: This is described by the Michaelis-Menten model for enzyme kinetics, which relates the initial velocity on an enzymatic reaction to the substrate concentration x trough the equation:

Marco Lattuada – Statistical and Numerical Methods for Chemical Engineers Nonlinear Regressions – Page # 3 Puromycin Kinetics The model:

Marco Lattuada – Statistical and Numerical Methods for Chemical Engineers Nonlinear Regressions – Page # 4 Model Linearization Puromycin Kinetics: Model Rearrangement: Linearized Model: Puromycin Kinetics: Model Rearrangement: Linearized Model:

Marco Lattuada – Statistical and Numerical Methods for Chemical Engineers Nonlinear Regressions – Page # 5 Model Linearization Regression Line  1 =  2 =

Marco Lattuada – Statistical and Numerical Methods for Chemical Engineers Nonlinear Regressions – Page # 6 Model Linearization Regression from linearized model  1 =  2 =

Marco Lattuada – Statistical and Numerical Methods for Chemical Engineers Nonlinear Regressions – Page # 7 Puromycin Kinetics The model:   Linearized model is needed to estimate  2

Marco Lattuada – Statistical and Numerical Methods for Chemical Engineers Nonlinear Regressions – Page # 8 Nonlinear Regression Object To minimize the objective function where n is the number of observations, y i the responses, x i is the vector of the observations,  the vector of the parameters and f(x i,  ) the nonlinear model function. It is possible to plot the objective function S(  ) as a function of the parameter values, in order to reveal the presence of a minimum. Object To minimize the objective function where n is the number of observations, y i the responses, x i is the vector of the observations,  the vector of the parameters and f(x i,  ) the nonlinear model function. It is possible to plot the objective function S(  ) as a function of the parameter values, in order to reveal the presence of a minimum.

Marco Lattuada – Statistical and Numerical Methods for Chemical Engineers Nonlinear Regressions – Page # 9 Objective Function S(  ) Contour plot of S(q) Minimum Estimated value of  from linearization

Marco Lattuada – Statistical and Numerical Methods for Chemical Engineers Nonlinear Regressions – Page # 10 Minimization of S(  ) Model linearization: where: so the residuals are: Search for minimum with Gauss-Newton method: Model linearization: where: so the residuals are: Search for minimum with Gauss-Newton method: J 0 = Jacobian

Marco Lattuada – Statistical and Numerical Methods for Chemical Engineers Nonlinear Regressions – Page # 11 Gauss-Newton Method Applied to S(  ) Convergence path of Gauss- Newton Method (  1 ) opt = (  2 ) opt =

Marco Lattuada – Statistical and Numerical Methods for Chemical Engineers Nonlinear Regressions – Page # 12 Nonlinear Regression Regression from linearized model

Marco Lattuada – Statistical and Numerical Methods for Chemical Engineers Nonlinear Regressions – Page # 13 Ellipsoidal Confidence Region The ellipsoidal confidence region can be evaluate from the linearized model around the point, which is the vector of the parameters for which the objective function has a minimum. In practice, every vector of the parameters  which satisfies the following condition: is within the confidence interval, where n is the number of observations, p the number of parameters and s the standard deviation: The ellipsoidal confidence region can be evaluate from the linearized model around the point, which is the vector of the parameters for which the objective function has a minimum. In practice, every vector of the parameters  which satisfies the following condition: is within the confidence interval, where n is the number of observations, p the number of parameters and s the standard deviation:

Marco Lattuada – Statistical and Numerical Methods for Chemical Engineers Nonlinear Regressions – Page # 14 Ellipsoidal Confidence Region 

Marco Lattuada – Statistical and Numerical Methods for Chemical Engineers Nonlinear Regressions – Page # 15 True Confidence Region for Parameters The real confidence region can be estimated by plotting the region of space for which:

Marco Lattuada – Statistical and Numerical Methods for Chemical Engineers Nonlinear Regressions – Page # 16 Matlab Nonlinear Regression Routine First, create a function providing the residuals for the n observation as a function of the parameter values: Then, use the routine 'nlinfit'; First, create a function providing the residuals for the n observation as a function of the parameter values: Then, use the routine 'nlinfit';

Marco Lattuada – Statistical and Numerical Methods for Chemical Engineers Nonlinear Regressions – Page # 17 Tukey-Ancombe Plot

Marco Lattuada – Statistical and Numerical Methods for Chemical Engineers Nonlinear Regressions – Page # 18 Normal Plot >> normplot(r)

Marco Lattuada – Statistical and Numerical Methods for Chemical Engineers Nonlinear Regressions – Page # 19 Matlab Estimation of Parameter CI Parameter confidence interval can be estimated by Matlab as follows: The confidence interval can be estimated using the following Matlab GUI: Parameter confidence interval can be estimated by Matlab as follows: The confidence interval can be estimated using the following Matlab GUI: