Bootstrap in refinement

Slides:



Advertisements
Similar presentations
Hypothesis testing and confidence intervals by resampling by J. Kárász.
Advertisements

Uncertainty in fall time surrogate Prediction variance vs. data sensitivity – Non-uniform noise – Example Uncertainty in fall time data Bootstrapping.
A Statistician’s Games * : Bootstrap, Bagging and Boosting * Please refer to “Game theory, on-line prediction and boosting” by Y. Freund and R. Schapire,
Estimating the reliability of a tree Reconstructed phylogenetic trees are almost certainly wrong. They are estimates of the true tree. But how reliable.
Confidence Intervals Underlying model: Unknown parameter We know how to calculate point estimates E.g. regression analysis But different data would change.
Resampling techniques Why resampling? Jacknife Cross-validation Bootstrap Examples of application of bootstrap.
Navigation Jeremy Wyatt School of Computer Science University of Birmingham.
2008 Chingchun 1 Bootstrap Chingchun Huang ( 黃敬群 ) Vision Lab, NCTU.
Bootstrapping LING 572 Fei Xia 1/31/06.
Examples of Ensemble Methods
Bootstrap Estimation of the Predictive Distributions of Reserves Using Paid and Incurred Claims Huijuan Liu Cass Business School Lloyd’s of London 11/07/2007.
Linear and generalised linear models Purpose of linear models Least-squares solution for linear models Analysis of diagnostics Exponential family and generalised.
Guide to Using Minitab 14 For Basic Statistical Applications To Accompany Business Statistics: A Decision Making Approach, 8th Ed. Chapter 7: Introduction.
2015 AprilUNIVERSITY OF HAIFA, DEPARTMENT OF STATISTICS, SEMINAR FOR M.A 1 Hastie, Tibshirani and Friedman.The Elements of Statistical Learning (2nd edition,
Bootstrap and Cross-Validation Bootstrap and Cross-Validation.
Biostatistics IV An introduction to bootstrap. 2 Getting something from nothing? In Rudolph Erich Raspe's tale, Baron Munchausen had, in one of his many.
Microeconometric Modeling William Greene Stern School of Business New York University.
8 Sampling Distribution of the Mean Chapter8 p Sampling Distributions Population mean and standard deviation,  and   unknown Maximal Likelihood.
Resampling techniques
Limits to Statistical Theory Bootstrap analysis ESM April 2006.
Lynn Lethbridge SHRUG November, What is Bootstrapping? A method to estimate a statistic’s sampling distribution Bootstrap samples are drawn repeatedly.
Association between genotype and phenotype
Bootstrap Event Study Tests Peter Westfall ISQS Dept. Joint work with Scott Hein, Finance.
Introduction to resampling in MATLAB. So you've done an experiment... Two independent datasets: control experimental Have n numbers in each dataset, representing.
[Topic 1-Regression] 1/37 1. Descriptive Tools, Regression, Panel Data.
CSCI 347, Data Mining Evaluation: Cross Validation, Holdout, Leave-One-Out Cross Validation and Bootstrapping, Sections 5.3 & 5.4, pages
How Good is a Model? How much information does AIC give us? –Model 1: 3124 –Model 2: 2932 –Model 3: 2968 –Model 4: 3204 –Model 5: 5436.
A bin-free Extended Maximum Likelihood Fit + Feldman-Cousins error analysis Peter Litchfield  A bin free Extended Maximum Likelihood method of fitting.
Logistic Regression Saed Sayad 1www.ismartsoft.com.
Project Plan Task 8 and VERSUS2 Installation problems Anatoly Myravyev and Anastasia Bundel, Hydrometcenter of Russia March 2010.
1/61: Topic 1.2 – Extensions of the Linear Regression Model Microeconometric Modeling William Greene Stern School of Business New York University New York.
Area Test for Observations Indexed by Time L. B. Green Middle Tennessee State University E. M. Boczko Vanderbilt University.
Quantifying Uncertainty
Bootstrapping James G. Anderson, Ph.D. Purdue University.
Overview G. Jogesh Babu. R Programming environment Introduction to R programming language R is an integrated suite of software facilities for data manipulation,
Data Science Credibility: Evaluating What’s Been Learned
Application of the Bootstrap Estimating a Population Mean
Chapter 13 – Ensembles and Uplift
How Good is a Model? How much information does AIC give us?
Lecture 19: Spatial Interpolation II
Microeconometric Modeling
R Data Manipulation Bootstrapping
Test for Mean of a Non-Normal Population – small n
Simulation: Sensitivity, Bootstrap, and Power
Quantifying uncertainty using the bootstrap
Bootstrap - Example Suppose we have an estimator of a parameter and we want to express its accuracy by its standard error but its sampling distribution.
Writing the executive summary section of your report
Introduction to Data Mining, 2nd Edition
You need: Pencil Agenda Scrap Paper AP log Math book Calculator
Particle filters for Robot Localization
Sampling Distribution
Sampling Distribution
BOOTSTRAPPING: LEARNING FROM THE SAMPLE
QQ Plot Quantile to Quantile Plot Quantile: QQ Plot:
Microeconometric Modeling
The Gamma PDF Eliason (1993).
Bootstrapping Jackknifing
Version 5.3 From SMILE string to dictionary (LIBCHECK): Now coot uses it Segment id is now used Automatic adjustment for weights Improved bond order extraction.
Cross-validation Brenda Thomson/ Peter Fox Data Analytics
Ensemble learning Reminder - Bagging of Trees Random Forest
Non-parametric Filters: Particle Filters
Bootstrap and randomization methods
Non-parametric Filters: Particle Filters
Microeconometric Modeling
A bootstrap method for estimators based on combined administrative and survey data Sander Scholtus (Statistics Netherlands) NTTS Conference 13 March 2019.
Introduction to Machine learning
Database for MR.
Fractional-Random-Weight Bootstrap
Bootstrapping and Bootstrapping Regression Models
Presentation transcript:

Bootstrap in refinement 28 March 2007 Bootstrap in refinement Gábor Bunkóczi

Computationally demanding! Bootstrap - basics Statistical method for estimating the sample distribution of an estimator. Procedure: Given an estimator (Ε) and a sample. Create a new sample by resampling the original sample WITH replacement. Calculate the estimator, store the value. Repeat 1-3 Nboot times (> 1000). Computationally demanding!

Bootstrap - aims Model validation Map improvement 1. R-factor distribution 2. Coordinate errors Map improvement 1. Bias removal 2. Resolution extension

Bootstrap - algorithm Sample: Resampling: Refinement: Accumulation: Fo-Fc normalised in each resolution shell by <│Fo-Fc│2> Resampling: 1. Generate ΔFnorm = (Fo-Fc) / Norm 2. Randomise: ΔFnorm → ΔFnorm, random 3. Calculate Fo = Fc + ΔFnorm, random * Norm Refinement: 1. Model randomisation 2. Refinement on “bootstrap” data 3. Calculate R/Rfree on original data Accumulation: 1. R-factors 2. Map coefficients

Bootstrap - implementation START: Model, Dataset Initial refinement to calculate normalisation factor Generate “bootstrap” datasets Refinement Refinement Refinement Extract data from log files Accumulate map coefficient END: R-factor distribution, multiple models

Bootstrap – results R-factors: Coordinates: distributions very tight further randomisation increases absolute value but does not make the distribution broader Coordinates:

Bootstrap – development 1. Resample residuals → resample likelihood P(Fo, Fc) 2. Resample ΔF → resample difference map 3. Improved normalisation