SlideShare una empresa de Scribd logo
1 de 20
Descargar para leer sin conexión
The Problem                      Univariate Analysis           Multivariate Analysis      Conclusion




                   How mathematicians predict the future?

                                 Instructor: Agnieszka Wyloma´ska
                                                             n

               Costanza Catalano, Angela Ciliberti, Gonçalo S. Matos, Allan S. Nielsen,
                                  Olga Polikarpova, Mattia Zanella

                            European Summer School in Industrial Mathematics
                                           Modelling Week


                                                   July 30, 2011



How mathematicians predict the future?                                                       ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis   Multivariate Analysis   Conclusion




Supplied Data for Analysis




How mathematicians predict the future?                                            ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis   Multivariate Analysis   Conclusion




Supplied Data for Analysis




How mathematicians predict the future?                                            ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis   Multivariate Analysis   Conclusion




Our Approach



                             Univariate Analysis
                                 Orstein-Unlenbeck Model
                                 Autoregressive Model
                             Multivariate Analysis
                                 Linear Regression




How mathematicians predict the future?                                            ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis       Multivariate Analysis   Conclusion



Orstein-Uhlenbeck process


Orstein-Uhlenbeck process

       Orstein-Uhlenbeck process
       The Orstein-Uhlenbeck process (or mean-reverting process) is
       defined by the following equation:

                                         dXt = θ(µ − Xt )dt + σdWt

       Where Wt is a Wiener process, t ∈ T ⊆ R+ represents time and
       θ > 0, µ and σ > 0 are time independent constants.

       Here Xt = log(St ) is the logarithm of the implied/nominal/real
       inflation St .

How mathematicians predict the future?                                                ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis   Multivariate Analysis   Conclusion



Orstein-Uhlenbeck process


Euler Maruyama method


       Euler Maruyama method
       The Euler Maruyama method is a method for the approximate
       numerical solution of a stochastic differential equation. In our
       case, for a partition of [t, t + 1] in n equal subintervals:

                                 Xn+1 = Xn + θ(µ − Xn )δ + σ∆Wn

       Where δ = 1/N is the length of the subintervals, and ∆Wn are
       independent identically distributed random varibles with expected
       value of 0 and a variance of δ.


How mathematicians predict the future?                                            ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis   Multivariate Analysis   Conclusion



Orstein-Uhlenbeck process


Empirical Distribution for 1 Step Prediction




How mathematicians predict the future?                                            ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis              Multivariate Analysis   Conclusion



AR(p)


Autoregressive model

        Autoregressive model
        The autoregressive model of order p, AR(p), is defined as:
                                                       p
                                         Yt = a0 +           ai Yt−i + εt
                                                       i=1

        Where a0 , a1 , . . . , ap are the parameters of the model and εt is
        independent identically distributed random variables.

        Here Yt = St − St−1 is the backward difference of the
        implied/nominal/real inflation St .

How mathematicians predict the future?                                                       ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis   Multivariate Analysis   Conclusion



AR(p)


Autocorrelation Function of the Implied Inflation




How mathematicians predict the future?                                            ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis   Multivariate Analysis   Conclusion



AR(p)


Forecast




How mathematicians predict the future?                                            ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis   Multivariate Analysis   Conclusion



AR(p)


Evolution of Probability Distributions




How mathematicians predict the future?                                            ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis   Multivariate Analysis   Conclusion



Confidence Bands


Confidence Band (close up)




How mathematicians predict the future?                                            ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis   Multivariate Analysis   Conclusion



Confidence Bands


Confidence Band (all view)




How mathematicians predict the future?                                            ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis   Multivariate Analysis   Conclusion



Confidence Bands


Confidence Band (2 Years Data)




How mathematicians predict the future?                                            ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis   Multivariate Analysis   Conclusion



Linear Regression


Correlation between Time Series




How mathematicians predict the future?                                            ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis                Multivariate Analysis   Conclusion



Linear Regression




       Linear Regression
       The multivariate regression model is:

                                                   Y = XT β + ε

                                                   E(Y) = XT β
                                                       ΣY = σ 2 1
       Where Y are the response variables and X are the explanatory
       variables.




How mathematicians predict the future?                                                         ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis   Multivariate Analysis   Conclusion



Linear Regression


Linear Regression Prediction




How mathematicians predict the future?                                            ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis   Multivariate Analysis   Conclusion



Linear Regression


Error in the Prediction




How mathematicians predict the future?                                            ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis   Multivariate Analysis   Conclusion



Linear Regression


Confidence Band




How mathematicians predict the future?                                            ESSIM
European Consortium for Mathematics in Industry
The Problem                      Univariate Analysis   Multivariate Analysis   Conclusion



Conclusion


Final Remarks




       • Summary:

                Confidence Band and Spread control
                Implied inflation, Real and Nominal seem to be correlated




How mathematicians predict the future?                                            ESSIM
European Consortium for Mathematics in Industry

Más contenido relacionado

Destacado

SAS Data Mining - Crime Modeling
SAS Data Mining - Crime ModelingSAS Data Mining - Crime Modeling
SAS Data Mining - Crime ModelingJohn Michael Croft
 
Risk evaluation for non-equity investments
Risk evaluation for non-equity investmentsRisk evaluation for non-equity investments
Risk evaluation for non-equity investmentsMattia Zanella
 
Multi variate presentation
Multi variate presentationMulti variate presentation
Multi variate presentationArun Kumar
 
Level of Measurement, Frequency Distribution,Stem & Leaf
Level of Measurement, Frequency Distribution,Stem & Leaf   Level of Measurement, Frequency Distribution,Stem & Leaf
Level of Measurement, Frequency Distribution,Stem & Leaf Qasim Raza
 
NFL 2013 Combine Data Multivariate Analysis
NFL 2013 Combine Data Multivariate AnalysisNFL 2013 Combine Data Multivariate Analysis
NFL 2013 Combine Data Multivariate AnalysisJohn Michael Croft
 
Modeling Online Hotel Choice: Conjoint analysis as a multivariate alternative...
Modeling Online Hotel Choice: Conjoint analysis as a multivariate alternative...Modeling Online Hotel Choice: Conjoint analysis as a multivariate alternative...
Modeling Online Hotel Choice: Conjoint analysis as a multivariate alternative...SKIM
 
Multivariate Data Analysis
Multivariate Data AnalysisMultivariate Data Analysis
Multivariate Data AnalysisMerul Romadhani
 
Multivariate data analysis
Multivariate data analysisMultivariate data analysis
Multivariate data analysisSetia Pramana
 
Multivariate Analysis An Overview
Multivariate Analysis An OverviewMultivariate Analysis An Overview
Multivariate Analysis An Overviewguest3311ed
 
Multivariate Analysis Techniques
Multivariate Analysis TechniquesMultivariate Analysis Techniques
Multivariate Analysis TechniquesMehul Gondaliya
 

Destacado (11)

SAS Data Mining - Crime Modeling
SAS Data Mining - Crime ModelingSAS Data Mining - Crime Modeling
SAS Data Mining - Crime Modeling
 
Risk evaluation for non-equity investments
Risk evaluation for non-equity investmentsRisk evaluation for non-equity investments
Risk evaluation for non-equity investments
 
Multi variate presentation
Multi variate presentationMulti variate presentation
Multi variate presentation
 
Level of Measurement, Frequency Distribution,Stem & Leaf
Level of Measurement, Frequency Distribution,Stem & Leaf   Level of Measurement, Frequency Distribution,Stem & Leaf
Level of Measurement, Frequency Distribution,Stem & Leaf
 
NFL 2013 Combine Data Multivariate Analysis
NFL 2013 Combine Data Multivariate AnalysisNFL 2013 Combine Data Multivariate Analysis
NFL 2013 Combine Data Multivariate Analysis
 
Modeling Online Hotel Choice: Conjoint analysis as a multivariate alternative...
Modeling Online Hotel Choice: Conjoint analysis as a multivariate alternative...Modeling Online Hotel Choice: Conjoint analysis as a multivariate alternative...
Modeling Online Hotel Choice: Conjoint analysis as a multivariate alternative...
 
Multivariate Data Analysis
Multivariate Data AnalysisMultivariate Data Analysis
Multivariate Data Analysis
 
Multivariate data analysis
Multivariate data analysisMultivariate data analysis
Multivariate data analysis
 
Multivariate analysis
Multivariate analysisMultivariate analysis
Multivariate analysis
 
Multivariate Analysis An Overview
Multivariate Analysis An OverviewMultivariate Analysis An Overview
Multivariate Analysis An Overview
 
Multivariate Analysis Techniques
Multivariate Analysis TechniquesMultivariate Analysis Techniques
Multivariate Analysis Techniques
 

Similar a How mathematicians predict the future?

2015-12-17 research seminar 2nd part
2015-12-17 research seminar 2nd part2015-12-17 research seminar 2nd part
2015-12-17 research seminar 2nd partifi8106tlu
 
5. Linear Algebra for Machine Learning: Singular Value Decomposition and Prin...
5. Linear Algebra for Machine Learning: Singular Value Decomposition and Prin...5. Linear Algebra for Machine Learning: Singular Value Decomposition and Prin...
5. Linear Algebra for Machine Learning: Singular Value Decomposition and Prin...Ceni Babaoglu, PhD
 
Lecture 1 test
Lecture 1 testLecture 1 test
Lecture 1 testfalcarragh
 
Cryptography Baby Step Giant Step
Cryptography Baby Step Giant StepCryptography Baby Step Giant Step
Cryptography Baby Step Giant StepSAUVIK BISWAS
 
A likelihood-free version of the stochastic approximation EM algorithm (SAEM)...
A likelihood-free version of the stochastic approximation EM algorithm (SAEM)...A likelihood-free version of the stochastic approximation EM algorithm (SAEM)...
A likelihood-free version of the stochastic approximation EM algorithm (SAEM)...Umberto Picchini
 
Tutorial on EM algorithm – Part 1
Tutorial on EM algorithm – Part 1Tutorial on EM algorithm – Part 1
Tutorial on EM algorithm – Part 1Loc Nguyen
 
05 history of cv a machine learning (theory) perspective on computer vision
05  history of cv a machine learning (theory) perspective on computer vision05  history of cv a machine learning (theory) perspective on computer vision
05 history of cv a machine learning (theory) perspective on computer visionzukun
 
Nature-Inspired Optimization Algorithms
Nature-Inspired Optimization Algorithms Nature-Inspired Optimization Algorithms
Nature-Inspired Optimization Algorithms Xin-She Yang
 
Jindřich Libovický - 2017 - Attention Strategies for Multi-Source Sequence-...
Jindřich Libovický - 2017 - Attention Strategies for Multi-Source Sequence-...Jindřich Libovický - 2017 - Attention Strategies for Multi-Source Sequence-...
Jindřich Libovický - 2017 - Attention Strategies for Multi-Source Sequence-...Association for Computational Linguistics
 
A Study on Performance Analysis of Different Prediction Techniques in Predict...
A Study on Performance Analysis of Different Prediction Techniques in Predict...A Study on Performance Analysis of Different Prediction Techniques in Predict...
A Study on Performance Analysis of Different Prediction Techniques in Predict...IJRES Journal
 
Markov chain monte_carlo_methods_for_machine_learning
Markov chain monte_carlo_methods_for_machine_learningMarkov chain monte_carlo_methods_for_machine_learning
Markov chain monte_carlo_methods_for_machine_learningAndres Mendez-Vazquez
 
Markov chain Monte Carlo methods and some attempts at parallelizing them
Markov chain Monte Carlo methods and some attempts at parallelizing themMarkov chain Monte Carlo methods and some attempts at parallelizing them
Markov chain Monte Carlo methods and some attempts at parallelizing themPierre Jacob
 
Numerical Solutions of Burgers' Equation Project Report
Numerical Solutions of Burgers' Equation Project ReportNumerical Solutions of Burgers' Equation Project Report
Numerical Solutions of Burgers' Equation Project ReportShikhar Agarwal
 
Computational methods for nanoscale bio sensors
Computational methods for nanoscale bio sensorsComputational methods for nanoscale bio sensors
Computational methods for nanoscale bio sensorsUniversity of Glasgow
 
Intro to Approximate Bayesian Computation (ABC)
Intro to Approximate Bayesian Computation (ABC)Intro to Approximate Bayesian Computation (ABC)
Intro to Approximate Bayesian Computation (ABC)Umberto Picchini
 
Econometrics of panel data - a presentation
Econometrics of panel data - a presentationEconometrics of panel data - a presentation
Econometrics of panel data - a presentationdrbojutjub
 

Similar a How mathematicians predict the future? (20)

2015-12-17 research seminar 2nd part
2015-12-17 research seminar 2nd part2015-12-17 research seminar 2nd part
2015-12-17 research seminar 2nd part
 
5. Linear Algebra for Machine Learning: Singular Value Decomposition and Prin...
5. Linear Algebra for Machine Learning: Singular Value Decomposition and Prin...5. Linear Algebra for Machine Learning: Singular Value Decomposition and Prin...
5. Linear Algebra for Machine Learning: Singular Value Decomposition and Prin...
 
Lecture 1 test
Lecture 1 testLecture 1 test
Lecture 1 test
 
Finite Difference Method for Nonlocal Singularly Perturbed Problem
Finite Difference Method for Nonlocal Singularly Perturbed ProblemFinite Difference Method for Nonlocal Singularly Perturbed Problem
Finite Difference Method for Nonlocal Singularly Perturbed Problem
 
Program on Mathematical and Statistical Methods for Climate and the Earth Sys...
Program on Mathematical and Statistical Methods for Climate and the Earth Sys...Program on Mathematical and Statistical Methods for Climate and the Earth Sys...
Program on Mathematical and Statistical Methods for Climate and the Earth Sys...
 
Cryptography Baby Step Giant Step
Cryptography Baby Step Giant StepCryptography Baby Step Giant Step
Cryptography Baby Step Giant Step
 
A likelihood-free version of the stochastic approximation EM algorithm (SAEM)...
A likelihood-free version of the stochastic approximation EM algorithm (SAEM)...A likelihood-free version of the stochastic approximation EM algorithm (SAEM)...
A likelihood-free version of the stochastic approximation EM algorithm (SAEM)...
 
Tutorial on EM algorithm – Part 1
Tutorial on EM algorithm – Part 1Tutorial on EM algorithm – Part 1
Tutorial on EM algorithm – Part 1
 
05 history of cv a machine learning (theory) perspective on computer vision
05  history of cv a machine learning (theory) perspective on computer vision05  history of cv a machine learning (theory) perspective on computer vision
05 history of cv a machine learning (theory) perspective on computer vision
 
Calculus
CalculusCalculus
Calculus
 
Nature-Inspired Optimization Algorithms
Nature-Inspired Optimization Algorithms Nature-Inspired Optimization Algorithms
Nature-Inspired Optimization Algorithms
 
Jindřich Libovický - 2017 - Attention Strategies for Multi-Source Sequence-...
Jindřich Libovický - 2017 - Attention Strategies for Multi-Source Sequence-...Jindřich Libovický - 2017 - Attention Strategies for Multi-Source Sequence-...
Jindřich Libovický - 2017 - Attention Strategies for Multi-Source Sequence-...
 
A Study on Performance Analysis of Different Prediction Techniques in Predict...
A Study on Performance Analysis of Different Prediction Techniques in Predict...A Study on Performance Analysis of Different Prediction Techniques in Predict...
A Study on Performance Analysis of Different Prediction Techniques in Predict...
 
MUMS Opening Workshop - Emulators for models and Complexity Reduction - Akil ...
MUMS Opening Workshop - Emulators for models and Complexity Reduction - Akil ...MUMS Opening Workshop - Emulators for models and Complexity Reduction - Akil ...
MUMS Opening Workshop - Emulators for models and Complexity Reduction - Akil ...
 
Markov chain monte_carlo_methods_for_machine_learning
Markov chain monte_carlo_methods_for_machine_learningMarkov chain monte_carlo_methods_for_machine_learning
Markov chain monte_carlo_methods_for_machine_learning
 
Markov chain Monte Carlo methods and some attempts at parallelizing them
Markov chain Monte Carlo methods and some attempts at parallelizing themMarkov chain Monte Carlo methods and some attempts at parallelizing them
Markov chain Monte Carlo methods and some attempts at parallelizing them
 
Numerical Solutions of Burgers' Equation Project Report
Numerical Solutions of Burgers' Equation Project ReportNumerical Solutions of Burgers' Equation Project Report
Numerical Solutions of Burgers' Equation Project Report
 
Computational methods for nanoscale bio sensors
Computational methods for nanoscale bio sensorsComputational methods for nanoscale bio sensors
Computational methods for nanoscale bio sensors
 
Intro to Approximate Bayesian Computation (ABC)
Intro to Approximate Bayesian Computation (ABC)Intro to Approximate Bayesian Computation (ABC)
Intro to Approximate Bayesian Computation (ABC)
 
Econometrics of panel data - a presentation
Econometrics of panel data - a presentationEconometrics of panel data - a presentation
Econometrics of panel data - a presentation
 

How mathematicians predict the future?

  • 1. The Problem Univariate Analysis Multivariate Analysis Conclusion How mathematicians predict the future? Instructor: Agnieszka Wyloma´ska n Costanza Catalano, Angela Ciliberti, Gonçalo S. Matos, Allan S. Nielsen, Olga Polikarpova, Mattia Zanella European Summer School in Industrial Mathematics Modelling Week July 30, 2011 How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 2. The Problem Univariate Analysis Multivariate Analysis Conclusion Supplied Data for Analysis How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 3. The Problem Univariate Analysis Multivariate Analysis Conclusion Supplied Data for Analysis How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 4. The Problem Univariate Analysis Multivariate Analysis Conclusion Our Approach Univariate Analysis Orstein-Unlenbeck Model Autoregressive Model Multivariate Analysis Linear Regression How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 5. The Problem Univariate Analysis Multivariate Analysis Conclusion Orstein-Uhlenbeck process Orstein-Uhlenbeck process Orstein-Uhlenbeck process The Orstein-Uhlenbeck process (or mean-reverting process) is defined by the following equation: dXt = θ(µ − Xt )dt + σdWt Where Wt is a Wiener process, t ∈ T ⊆ R+ represents time and θ > 0, µ and σ > 0 are time independent constants. Here Xt = log(St ) is the logarithm of the implied/nominal/real inflation St . How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 6. The Problem Univariate Analysis Multivariate Analysis Conclusion Orstein-Uhlenbeck process Euler Maruyama method Euler Maruyama method The Euler Maruyama method is a method for the approximate numerical solution of a stochastic differential equation. In our case, for a partition of [t, t + 1] in n equal subintervals: Xn+1 = Xn + θ(µ − Xn )δ + σ∆Wn Where δ = 1/N is the length of the subintervals, and ∆Wn are independent identically distributed random varibles with expected value of 0 and a variance of δ. How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 7. The Problem Univariate Analysis Multivariate Analysis Conclusion Orstein-Uhlenbeck process Empirical Distribution for 1 Step Prediction How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 8. The Problem Univariate Analysis Multivariate Analysis Conclusion AR(p) Autoregressive model Autoregressive model The autoregressive model of order p, AR(p), is defined as: p Yt = a0 + ai Yt−i + εt i=1 Where a0 , a1 , . . . , ap are the parameters of the model and εt is independent identically distributed random variables. Here Yt = St − St−1 is the backward difference of the implied/nominal/real inflation St . How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 9. The Problem Univariate Analysis Multivariate Analysis Conclusion AR(p) Autocorrelation Function of the Implied Inflation How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 10. The Problem Univariate Analysis Multivariate Analysis Conclusion AR(p) Forecast How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 11. The Problem Univariate Analysis Multivariate Analysis Conclusion AR(p) Evolution of Probability Distributions How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 12. The Problem Univariate Analysis Multivariate Analysis Conclusion Confidence Bands Confidence Band (close up) How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 13. The Problem Univariate Analysis Multivariate Analysis Conclusion Confidence Bands Confidence Band (all view) How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 14. The Problem Univariate Analysis Multivariate Analysis Conclusion Confidence Bands Confidence Band (2 Years Data) How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 15. The Problem Univariate Analysis Multivariate Analysis Conclusion Linear Regression Correlation between Time Series How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 16. The Problem Univariate Analysis Multivariate Analysis Conclusion Linear Regression Linear Regression The multivariate regression model is: Y = XT β + ε E(Y) = XT β ΣY = σ 2 1 Where Y are the response variables and X are the explanatory variables. How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 17. The Problem Univariate Analysis Multivariate Analysis Conclusion Linear Regression Linear Regression Prediction How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 18. The Problem Univariate Analysis Multivariate Analysis Conclusion Linear Regression Error in the Prediction How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 19. The Problem Univariate Analysis Multivariate Analysis Conclusion Linear Regression Confidence Band How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry
  • 20. The Problem Univariate Analysis Multivariate Analysis Conclusion Conclusion Final Remarks • Summary: Confidence Band and Spread control Implied inflation, Real and Nominal seem to be correlated How mathematicians predict the future? ESSIM European Consortium for Mathematics in Industry