The KOSMOSHOWS What is it ? The statistic inside What it can do ? Future development Demonstration A. Tilquin (CPPM)

Slides:



Advertisements
Similar presentations
Introduction to Data Assimilation NCEO Data-assimilation training days 5-7 July 2010 Peter Jan van Leeuwen Data Assimilation Research Center (DARC) University.
Advertisements

Modeling of Data. Basic Bayes theorem Bayes theorem relates the conditional probabilities of two events A, and B: A might be a hypothesis and B might.
Design of Experiments Lecture I
Multilevel analysis with EQS. Castello2004 Data is datamlevel.xls, datamlevel.sav, datamlevel.ess.
Factorial Mixture of Gaussians and the Marginal Independence Model Ricardo Silva Joint work-in-progress with Zoubin Ghahramani.
CmpE 104 SOFTWARE STATISTICAL TOOLS & METHODS MEASURING & ESTIMATING SOFTWARE SIZE AND RESOURCE & SCHEDULE ESTIMATING.
Dynamics of Learning VQ and Neural Gas Aree Witoelar, Michael Biehl Mathematics and Computing Science University of Groningen, Netherlands in collaboration.
Uncertainty Representation. Gaussian Distribution variance Standard deviation.
Visual Recognition Tutorial
Artificial Learning Approaches for Multi-target Tracking Jesse McCrosky Nikki Hu.
Prénom Nom Document Analysis: Parameter Estimation for Pattern Recognition Prof. Rolf Ingold, University of Fribourg Master course, spring semester 2008.
3D Geometry for Computer Graphics. 2 The plan today Least squares approach  General / Polynomial fitting  Linear systems of equations  Local polynomial.
G. Cowan Lectures on Statistical Data Analysis 1 Statistical Data Analysis: Lecture 10 1Probability, Bayes’ theorem, random variables, pdfs 2Functions.
CF-3 Bank Hapoalim Jun-2001 Zvi Wiener Computational Finance.
Maximum likelihood Conditional distribution and likelihood Maximum likelihood estimations Information in the data and likelihood Observed and Fisher’s.
Empirical Saddlepoint Approximations for Statistical Inference Fallaw Sowell Tepper School of Business Carnegie Mellon University September 2006.
Linear Regression Models Based on Chapter 3 of Hastie, Tibshirani and Friedman Slides by David Madigan.
7. Least squares 7.1 Method of least squares K. Desch – Statistical methods of data analysis SS10 Another important method to estimate parameters Connection.
Maximum likelihood (ML)
Principles of Least Squares
Objectives of Multiple Regression
EE513 Audio Signals and Systems Statistical Pattern Classification Kevin D. Donohue Electrical and Computer Engineering University of Kentucky.
Weak Lensing 3 Tom Kitching. Introduction Scope of the lecture Power Spectra of weak lensing Statistics.
Model Inference and Averaging
G. Cowan Lectures on Statistical Data Analysis Lecture 3 page 1 Lecture 3 1 Probability (90 min.) Definition, Bayes’ theorem, probability densities and.
A Neural Network MonteCarlo approach to nucleon Form Factors parametrization Paris, ° CLAS12 Europen Workshop In collaboration with: A. Bacchetta.
1.3 The Intersection Point of Lines System of Equation A system of two equations in two variables looks like: – Notice, these are both lines. Linear Systems.
Using Neural Networks to Predict Claim Duration in the Presence of Right Censoring and Covariates David Speights Senior Research Statistician HNC Insurance.
CS 782 – Machine Learning Lecture 4 Linear Models for Classification  Probabilistic generative models  Probabilistic discriminative models.
Parameter estimation. 2D homography Given a set of (x i,x i ’), compute H (x i ’=Hx i ) 3D to 2D camera projection Given a set of (X i,x i ), compute.
RSVM: Reduced Support Vector Machines Y.-J. Lee & O. L. Mangasarian First SIAM International Conference on Data Mining Chicago, April 6, 2001 University.
Chapter 7 Point Estimation of Parameters. Learning Objectives Explain the general concepts of estimating Explain important properties of point estimators.
BOĞAZİÇİ UNIVERSITY DEPARTMENT OF MANAGEMENT INFORMATION SYSTEMS MATLAB AS A DATA MINING ENVIRONMENT.
Data Modeling Patrice Koehl Department of Biological Sciences National University of Singapore
1  The Problem: Consider a two class task with ω 1, ω 2   LINEAR CLASSIFIERS.
V0 analytical selection Marian Ivanov, Alexander Kalweit.
Lecture 2: Statistical learning primer for biologists
OPERATING SYSTEMS CS 3530 Summer 2014 Systems and Models Chapter 03.
Machine Learning 5. Parametric Methods.
Chapter 2-OPTIMIZATION G.Anuradha. Contents Derivative-based Optimization –Descent Methods –The Method of Steepest Descent –Classical Newton’s Method.
Regression Analysis1. 2 INTRODUCTION TO EMPIRICAL MODELS LEAST SQUARES ESTIMATION OF THE PARAMETERS PROPERTIES OF THE LEAST SQUARES ESTIMATORS AND ESTIMATION.
1 Chapter 8: Model Inference and Averaging Presented by Hui Fang.
The Unscented Particle Filter 2000/09/29 이 시은. Introduction Filtering –estimate the states(parameters or hidden variable) as a set of observations becomes.
Review of statistical modeling and probability theory Alan Moses ML4bio.
Maximum likelihood estimators Example: Random data X i drawn from a Poisson distribution with unknown  We want to determine  For any assumed value of.
University of Colorado Boulder ASEN 5070 Statistical Orbit determination I Fall 2012 Professor George H. Born Professor Jeffrey S. Parker Lecture 9: Least.
BlueFin Best Linear Unbiased Estimate Fisher Information aNalysis Andrea Valassi (IT-SDC) based on the work done with Roberto Chierici TOPLHCWG meeting.
G. Cowan Lectures on Statistical Data Analysis Lecture 10 page 1 Statistical Data Analysis: Lecture 10 1Probability, Bayes’ theorem 2Random variables and.
Bayesian Brain Probabilistic Approaches to Neural Coding 1.1 A Probability Primer Bayesian Brain Probabilistic Approaches to Neural Coding 1.1 A Probability.
Statistics 350 Lecture 2. Today Last Day: Section Today: Section 1.6 Homework #1: Chapter 1 Problems (page 33-38): 2, 5, 6, 7, 22, 26, 33, 34,
Environmental Data Analysis with MatLab 2 nd Edition Lecture 22: Linear Approximations and Non Linear Least Squares.
Chapter 12 Case Studies Part B. Control System Design.
Data Modeling Patrice Koehl Department of Biological Sciences
Overview Modern chip designs have multiple IP components with different process, voltage, temperature sensitivities Optimizing mix to different customer.
OPERATING SYSTEMS CS 3502 Fall 2017
Deep Feedforward Networks
Probability Theory and Parameter Estimation I
CH 5: Multivariate Methods
Statistical Methods For Engineers
Linear Systems.
Ying shen Sse, tongji university Sep. 2016
Lecture 2 – Monte Carlo method in finance
Where did we stop? The Bayes decision rule guarantees an optimal classification… … But it requires the knowledge of P(ci|x) (or p(x|ci) and P(ci)) We.
Lecture 5 Unsupervised Learning in fully Observed Directed and Undirected Graphical Models.
Lecture 3 1 Probability Definition, Bayes’ theorem, probability densities and their properties, catalogue of pdfs, Monte Carlo 2 Statistical tests general.
EE513 Audio Signals and Systems
6.1 Introduction to Chi-Square Space
Computing and Statistical Data Analysis / Stat 7
#21 Marginalize vs. Condition Uninteresting Fitted Parameters
Parametric Methods Berlin Chen, 2005 References:
Presentation transcript:

The KOSMOSHOWS What is it ? The statistic inside What it can do ? Future development Demonstration A. Tilquin (CPPM)

What is it? Interactive graphic tools for cosmological parameters extractions Now only for SN experiments It is written in IDL and can be run on Linux or Windows It can be used as a physics analysis tools Fitting of parameters Errors and full correlation computation Ellipse construction It can be used as a cosmological simulator Standard cosmology User define cosmology

The statistic We have to minimized %  k According to a set of measurement and a given parametric model, what are the best parameter values ? The most probable. Use of the likelihood We have to maximized L %  k  If we take the log: 

Minimization At the first order we use an iterative equation to solve If the model is linear, minimum is found in one step

Error computation By construction the covariance matrix is the second derivative of the chi2 at the minimum This is usually called the Fisher analysis. It gives the correct answer if the model is linear. The error are independent of the measured points. In such an analysis errors are symmetric.

Correct error computation If the model is non linear the Fisher analysis is an approximation. To get the correct errors we have to solve the equation: Where s 2 =1 for a 1  error (68% probability). They are 2 solutions,  +,  -. Fisher analysis  2 =  min 2 +1 -- Second minimum ++ Asymmetric errors

Contour construction The Fisher approach: Quadratic form  Ellipse For non linear model the  2 is no more a quadratic from: 1.Integration over unwanted variables (average contour): 2.Minimization over unwanted variables (most probable contour): Unwanted variables should be marginalized :

Contours —Fisher —Rigorous 39% —Fisher —Rigorous M s,  m,  X fitM s,  m,  X,w 0 fit M s,  m,  X,w 0 w 1 fit

Monte Carlo verification Ellipse:(30.4  2)% Contour:(39.6  2)% 1  contour = 39.3%

Limitation of the method P(ellipse) = (49.5  1.6)% P(contour)=(54.7  1.6)% The wrong normalization is due to the unphysical red zone ->complex luminosity distance !!! Solved in the kosmoshows by taking the real part of of luminosity distance.

What it can do 1.Fit real data or simulated data with various cosmology 2.Estimate error and correlation 3.Use external constraint (prior) 4.Construct probability contour 5.Perform statistical analysis as pool or probability  2 6.Perform evolution of errors and central values with respect to parameters 7.Simulate user defined cosmology and experiment 8.Produce various plots and save them in postscript 9.Read external data, write ntuple to be read by a graphic tool (paw) 10.Call external procedure to perform specific user analysis 11.Etc…

Future development Make this tools more “professional” Make it public. New language (JAVA/ROOT) ? Perform combined fit by using directly the public software: SN1a/CMB/Weak lensing/Baryon oscillation/GRB….. Because of CMB complexity software (1000 hours of CPU to construct contour) new developments are necessary Parallelism on PC cluster Use of Neural Network to speed up processing. Parameterization of the  2 with a NN

Now the demonstration Usually the most difficult part of an online talk Sorry if program crashes But it is still under development lines of code !!!!