Stat Seminar 10/26/06. General position results on uniqueness of nonrandomized Group-sequential decision procedures in Clinical Trials Eric Slud, UMCP

Size: px
Start display at page:

Download "Stat Seminar 10/26/06. General position results on uniqueness of nonrandomized Group-sequential decision procedures in Clinical Trials Eric Slud, UMCP"

Transcription

1 Stat Seminar 10/26/06 General position results on uniqueness of nonrandomized Group-sequential decision procedures in Clinical Trials Eric Slud, UMCP OUTLINE I. Two-Sample Clinical Trial Statistics A. Large-sample background B. Loss-functions & constraints on actions II. Decision Theory Formulation III. General Two-look Optimized Plans A. Reductions & results B. Computed Boundary Example C. Role of uniqueness & nonrandomized solutions IV. Random Perturbations of Loss Functions A. General result on a.s. unique minima B. Proof Sketch: a.s. unique optimal plans Joint work with Eric Leifer, NHLBI. 1

2 Two-Sample Clinical Trial Statistics Data format : (E i,ti, i,z i, for analysis at times t. i =1,...,N A (τ)) E i entry-times, N A arrival-counting, τ accrual-horizon X i failure time, C i indep. right-cens., Z i trt. gp. (X i,c i cond. indep. given Z i & strat. variable V i ) T i = X i C i (t E i ), i = I [Xi C i (t E i ]) PROBLEM: test H 0 : S X Z (t z) S X (t),z=0, 1 with multiple interim looks & experimentwise validity. TEST STATISTIC : for look at t, define Y z (s) = j M(t )= i I [Zj =z,t j s], Y (s) =Y 1 (s)+y 0 (s) at-risk K(s, Ŝ X (s)){z i Y 1 (T i s) Y (T i s) } i di [T i s] asympt. indep. incr., with estimated variance ˆV (t ) = i K 2 (s, ŜX(s)){ Y 1 (T i s)y0 (T i s) Y (Ti } i s) di [T i s] Reject if M(t ) ˆV (t ) b(t ), Accept if a(t ) 2

3 Abstracted Asymptotic Problem Most methods rely on asymptotically Gaussian timeindexed statistic-numerator n 1/2 M(t) with indep. incr. s, and variance function V (t) to be estimated in real or information time. V (t) in H 0 is functional of ( S X,S C Z, Λ A E(N A )) K 2 (S X (s)) p(1 p)λ A(t s)s C Z (s 1)S C Z (s 0) Λ A (t)(p(s C Z (s 1) + (1 p)s C Z (s 0)) df X(s) Control parameters: At each t = t k, can choose t k+1,a k+1,b k+1 PROBLEM: To optimize expected Loss or Cost over times and cutoffs while maintaining overall nominal significance level. Main Computational Method of optimizing boundaries is parametric search for parametric boundary classes, or backward induction. 3

4 Decision-Theory Ingredients Actions: look-times t k and boundaries b k = b(t ) for W/ ˆV 1/2 for now consider only upper rejection boundary Prior: π(ϑ) density for group-difference log hazard ratio parameter ϑ (within semiparametric model). Losses: costs of experimentation c 1 (t, ϑ), wrong decision c 2 (ϑ), late correct decision c 3 (t, ϑ), these loss elements introduced in Leifer (2000) thesis. Costs are economic within-trial, ethical within-trial, and economic after-trial. In 2-Look Trials, adaptive actions are: t 1 constant depending on prior & losses Followup time t 2 = t 2 (W (t 1 ),t 1 ) which if = t 1 indicates imediate Reject or Accept, and final boundary: reject iff W (t 2 ) b(t 2,t 1,W(t 1 )) 4

5 More on Decision Theory Problem (I) state-of-nature parameter ϑ R, target parameter z = γ(ϑ) (= ϑ in estimation, I[ϑ 0] in testing), and prior prob. π on Θ (II) action-space A = {(t,a):t =(t 1,...,t K ), 0 t 1 t K T, a γ(θ)} (III) observable process W =(W (x), 0 x T ) Wiener process with drift ϑ, and suff. stat. W (τ) for data up to stopping-time τ, and auxiliary randomization r.v. U Unif[0, 1] independent of W (IV) strategies consist of: increasing nonnegative stoppingtime random variables τ =(τ 1,...,τ K ) 0 τ 1 τ 2 τ K T a.s. satisfying causality restrictions on measurability, [τ j x] σ(τ i,i<j; W (u), u x ; U), and terminal action r.v. a γ(θ) meas. wrt (W (x), x t K ) and U 5

6 (V) a loss-function L(t,a,ϑ,λ) depending on looktimes t only as smooth fcn of terminal time t K, with Lagrange multipliers λ 0,λ 1 L(t,a,ϑ,λ) = C(t K,a,z)+λ 0 aδ ϑ,0 + λ 1 (1 a) δ ϑ,ϑ1 where z = I [ϑ 0], ϑ 1 > 0 and C(t, a, z) = c 1 (t, ϑ) +I [z a] c 2 (ϑ) +I [z=a] c 3 (t, ϑ) and c j (,ϑ), c 3 (t, ϑ) c 2 (ϑ). (VI) risk function to be minimized over d =(τ,a) is r(d, λ) = Θ E ϑ L(τ, α, ϑ, λ) dπ(ϑ) (1) NOTE: For terminal time τ K = s and data history (W (u), u s), there exists unique optimal terminal action I [W (s) w(s)], with w( ) = w(,λ 0,λ 1 ) uniquely determined by e θw(y) θ 2 y/2 g 1 (y,ϑ,λ 0,λ 1 ) dπ(θ) = 0 where g 1 (y) =(c 2 (θ) c 3 (y,θ)) (2I [θ 0] 1)+ λ 0 π 0 I [θ=0] λ 1 π 1 I [θ=θ1 ] 6

7 Optimizing the Decision Rule We know that at the final look-time, the best decision is to Reject when W (s) w(s), w computed uniquely from s, λ 0,λ 1. Now let t 1 be fixed and imagine W (t 1 )/ t 1 = x observed. Can express conditional expected final loss at time t 1 + s in form r 2 (t 1,x,s) R(t 1,x,s,ϑ) dπ(ϑ) = e θx t 1 θ 2 t 1 /2 {a 0 (t 1 + s, ϑ) +a 1 (t 1 + s, ϑ) 1 Φ w(t 1 + s) x t 1 s ϑ s } dπ(θ) where a 0 (t, ϑ) = c 1 (t, ϑ)+c 2 (ϑ) I [ϑ>0] + λ 1 π 1 I [θ=θ1 ] + c 3 (t, ϑ) I [ϑ 0] Backward Induction Idea: if we can uniquely optimize r 2 (t 1,x,s) in s = s(t 1,x) uniquely (a.e. x), and then choose t 1 uniquely to minimize 1 R(t1,x, s(t 1,x), ϑ) e (x ϑ t 1 ) 2 /2 dx dπ(ϑ) 2π 7

8 Optimizing, Continued then we would have specified the unique nonrandomized optimal decision rule through the constant t 1 and functions s(t 1,x), w(t 1 + s). But for these general loss functions there is no hope of explicit formulas, and even showing unique optima is complicated by the form of optimal s which will generally be 0 on the complement of some interval. This behavior is unavoidable, reflecting the necessity to stop after one look at the data at time t 1 if the observed statistic value W (t 1 ) is too extreme! We show this next is a computed example. 8

9 Optimal Boundary in 2-Look Example Data: W (t) =B(t i )+ϑt i, i =1, 2 t 1 is fixed in advance, continuation-time t 2 t 1 0 is chosen as function of W (t 1 ). Loss for stopping at τ with Rejection indicator z : c 1 (τ,ϑ) +c 3 (τ,ϑ) +z (c 2 (ϑ) c 3 (τ,ϑ)) (2I [ϑ 0] 1) Problem to find min-risk test under prior π(dϑ), with sig. level α and type II error at ϑ 1 β. Under regularity conditions on loss elements (piecewise smoothness, c 2 c 3, c 1 ) and prior π(dϑ) assigning positive mass to neighborhoods of 0, ϑ 1 > 0: can show that optimal procedures are nonrandomized (w.p.1 after smallrandom perturbation of c 1 ) and unique, rejecting for W (t 2 ) b 2 (t 1,t 2,W(t 1 )) V (t 2 ). Example. α =.025, β =.1, ϑ 1 = log(1.5), time scaled so τ fix = 1. Optimized t 1 =.42 τ fix. e ϑ = hazard ratio π({ϑ}) c 1 (t, ϑ) t t t t t c 2 (ϑ)

10 Total Trial Time normalized first-look statistic U 1

11 Second Look Critical Value normalized first-look statistic U 1

12 = -0.2 = 0.4 = 0.56 (lower bdy of cont. reg.) s s s = 0.7 =1 = s s s = 2.56 (upper bdy of cont. reg.) = 2.7 = s s s

13 Lemma on Unique Random-Function Minima Lemma 1 (extending Bulinskaya 1961 ) Let K be a fixed compact interval in R, and suppose that g 1,g 2 are fixed smooth real-valued functions on K, with inf K g 2 > 0. Let ζ( ) =ζ(,ω) be a random process on K with sample paths a.s. in C 2 (K) such that for all δ>0, the joint density of (ζ(s 1 ),ζ (s 1 ),ζ(s 2 ),ζ (s 2 )) (with respect to Lebesgue measure, on R 4 ) is uniformly bounded, uniformly over all s 1,s 2 K for which s 1 s 2 δ. Then (i) With probability 1 (ω), there do not exist distinct elements s 1,s 2 K such that the sample path of the function ρ(s) g 1 (s) +g 2 (s)ζ(s) satisfies ρ (s 1 )=ρ (s 2 )=0 along with ρ(s 1 )=ρ(s 2 ), and (ii) Under the assumptions above, with probability 1, the number of zeroes of ρ on K is finite, and ρ (s) 0 at all values of s for which ρ (s) =0. Corollary. Analogous result for minima wrt s with functions g j (t 1,x,s) holds a.e. in (t 1,x,ω). Via Implicit Function Theorem for minimizer s = s(t 1,x) find s a.e. smooth in its arguments, a.s. for argument ω. 14

14 Further Steps to Establish a.s. unique rule Want to substitute s = s(t 1,x) and integrate out standardized variable x = W (t 1 )/ t 1 to get smooth time-t 1 risk r 1 (t 1 ) except that this function s is only a.e. defined and smooth. Introduce technical assumption for small ɛ>0 that t 2 t 1 + ɛ whenever t 2 >t 1 (No real loss in generality: can likely prove it directly.) Remaining proof-step to prove r 1 (t 1 ) smooth: for V Unif[0,ɛ) indep. of other data, r 1 (t 1 ) = min (E(r 2 (t 1,W(t 1 )/ t 1, 0+), E( r 2 (t 1 + V, W (t 1 + V ) t1 + V,s(t 1 + V, W (t 1 + V ) t1 + V ))) Last expectations with respect to measure on (ϑ, W ( )) given by A P ϑ (dw )dπ(ϑ). Overall results hold a.s. (ω) for a.e. (λ 0,λ 1 ). But that seems to be enough. 15

15 References For clinical-trial large sample theory: Tsiatis (1982), Andersen & Gill (1982), Sellke & Siegmund (1983), Slud (1984) For group-sequential boundaries: Pocock (1977), O Brien & Fleming (1979), Slud & Wei (1982), Lan & DeMets (1983), Pampallona & Tsiatis (1994) plus others For adaptive modifications: Burman & Sondesson (2006, current Biometrics) For decision-theoretic formulations: Anscombe (1963), Colton (1963), Berger (1985 book), Siegmund (1985 book), Jennison (1987), Freedman & Spiegelhalter (1989), Leifer (2000, thesis) For level-crossings theorems: Bulinskaya (1962) cited in Cramer and Leadbetter (1967) Stationary and Related Stochastic Processes book 16

Independent Increments in Group Sequential Tests: A Review

Independent Increments in Group Sequential Tests: A Review Independent Increments in Group Sequential Tests: A Review KyungMann Kim kmkim@biostat.wisc.edu University of Wisconsin-Madison, Madison, WI, USA July 13, 2013 Outline Early Sequential Analysis Independent

More information

Testing a secondary endpoint after a group sequential test. Chris Jennison. 9th Annual Adaptive Designs in Clinical Trials

Testing a secondary endpoint after a group sequential test. Chris Jennison. 9th Annual Adaptive Designs in Clinical Trials Testing a secondary endpoint after a group sequential test Christopher Jennison Department of Mathematical Sciences, University of Bath, UK http://people.bath.ac.uk/mascj 9th Annual Adaptive Designs in

More information

The Design of a Survival Study

The Design of a Survival Study The Design of a Survival Study The design of survival studies are usually based on the logrank test, and sometimes assumes the exponential distribution. As in standard designs, the power depends on The

More information

Optimising Group Sequential Designs. Decision Theory, Dynamic Programming. and Optimal Stopping

Optimising Group Sequential Designs. Decision Theory, Dynamic Programming. and Optimal Stopping : Decision Theory, Dynamic Programming and Optimal Stopping Christopher Jennison Department of Mathematical Sciences, University of Bath, UK http://people.bath.ac.uk/mascj InSPiRe Conference on Methodology

More information

Published online: 10 Apr 2012.

Published online: 10 Apr 2012. This article was downloaded by: Columbia University] On: 23 March 215, At: 12:7 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 172954 Registered office: Mortimer

More information

Biometrika Trust. Biometrika Trust is collaborating with JSTOR to digitize, preserve and extend access to Biometrika.

Biometrika Trust. Biometrika Trust is collaborating with JSTOR to digitize, preserve and extend access to Biometrika. Biometrika Trust Discrete Sequential Boundaries for Clinical Trials Author(s): K. K. Gordon Lan and David L. DeMets Reviewed work(s): Source: Biometrika, Vol. 70, No. 3 (Dec., 1983), pp. 659-663 Published

More information

An Adaptive Futility Monitoring Method with Time-Varying Conditional Power Boundary

An Adaptive Futility Monitoring Method with Time-Varying Conditional Power Boundary An Adaptive Futility Monitoring Method with Time-Varying Conditional Power Boundary Ying Zhang and William R. Clarke Department of Biostatistics, University of Iowa 200 Hawkins Dr. C-22 GH, Iowa City,

More information

Pubh 8482: Sequential Analysis

Pubh 8482: Sequential Analysis Pubh 8482: Sequential Analysis Joseph S. Koopmeiners Division of Biostatistics University of Minnesota Week 5 Course Summary So far, we have discussed Group sequential procedures for two-sided tests Group

More information

Convergence of Quantum Statistical Experiments

Convergence of Quantum Statistical Experiments Convergence of Quantum Statistical Experiments M!d!lin Gu"! University of Nottingham Jonas Kahn (Paris-Sud) Richard Gill (Leiden) Anna Jen#ová (Bratislava) Statistical experiments and statistical decision

More information

Efficient Semiparametric Estimators via Modified Profile Likelihood in Frailty & Accelerated-Failure Models

Efficient Semiparametric Estimators via Modified Profile Likelihood in Frailty & Accelerated-Failure Models NIH Talk, September 03 Efficient Semiparametric Estimators via Modified Profile Likelihood in Frailty & Accelerated-Failure Models Eric Slud, Math Dept, Univ of Maryland Ongoing joint project with Ilia

More information

1 Simulating normal (Gaussian) rvs with applications to simulating Brownian motion and geometric Brownian motion in one and two dimensions

1 Simulating normal (Gaussian) rvs with applications to simulating Brownian motion and geometric Brownian motion in one and two dimensions Copyright c 2007 by Karl Sigman 1 Simulating normal Gaussian rvs with applications to simulating Brownian motion and geometric Brownian motion in one and two dimensions Fundamental to many applications

More information

Sample size re-estimation in clinical trials. Dealing with those unknowns. Chris Jennison. University of Kyoto, January 2018

Sample size re-estimation in clinical trials. Dealing with those unknowns. Chris Jennison. University of Kyoto, January 2018 Sample Size Re-estimation in Clinical Trials: Dealing with those unknowns Christopher Jennison Department of Mathematical Sciences, University of Bath, UK http://people.bath.ac.uk/mascj University of Kyoto,

More information

Interim Monitoring of Clinical Trials: Decision Theory, Dynamic Programming. and Optimal Stopping

Interim Monitoring of Clinical Trials: Decision Theory, Dynamic Programming. and Optimal Stopping Interim Monitoring of Clinical Trials: Decision Theory, Dynamic Programming and Optimal Stopping Christopher Jennison Department of Mathematical Sciences, University of Bath, UK http://people.bath.ac.uk/mascj

More information

Monitoring clinical trial outcomes with delayed response: incorporating pipeline data in group sequential designs. Christopher Jennison

Monitoring clinical trial outcomes with delayed response: incorporating pipeline data in group sequential designs. Christopher Jennison Monitoring clinical trial outcomes with delayed response: incorporating pipeline data in group sequential designs Christopher Jennison Department of Mathematical Sciences, University of Bath http://people.bath.ac.uk/mascj

More information

Stochastic Differential Equations.

Stochastic Differential Equations. Chapter 3 Stochastic Differential Equations. 3.1 Existence and Uniqueness. One of the ways of constructing a Diffusion process is to solve the stochastic differential equation dx(t) = σ(t, x(t)) dβ(t)

More information

PROBABILITY: LIMIT THEOREMS II, SPRING HOMEWORK PROBLEMS

PROBABILITY: LIMIT THEOREMS II, SPRING HOMEWORK PROBLEMS PROBABILITY: LIMIT THEOREMS II, SPRING 15. HOMEWORK PROBLEMS PROF. YURI BAKHTIN Instructions. You are allowed to work on solutions in groups, but you are required to write up solutions on your own. Please

More information

On the stochastic nonlinear Schrödinger equation

On the stochastic nonlinear Schrödinger equation On the stochastic nonlinear Schrödinger equation Annie Millet collaboration with Z. Brzezniak SAMM, Paris 1 and PMA Workshop Women in Applied Mathematics, Heraklion - May 3 211 Outline 1 The NL Shrödinger

More information

Brownian Motion. 1 Definition Brownian Motion Wiener measure... 3

Brownian Motion. 1 Definition Brownian Motion Wiener measure... 3 Brownian Motion Contents 1 Definition 2 1.1 Brownian Motion................................. 2 1.2 Wiener measure.................................. 3 2 Construction 4 2.1 Gaussian process.................................

More information

Closest Moment Estimation under General Conditions

Closest Moment Estimation under General Conditions Closest Moment Estimation under General Conditions Chirok Han and Robert de Jong January 28, 2002 Abstract This paper considers Closest Moment (CM) estimation with a general distance function, and avoids

More information

Pubh 8482: Sequential Analysis

Pubh 8482: Sequential Analysis Pubh 8482: Sequential Analysis Joseph S. Koopmeiners Division of Biostatistics University of Minnesota Week 8 P-values When reporting results, we usually report p-values in place of reporting whether or

More information

Machine Learning. Theory of Classification and Nonparametric Classifier. Lecture 2, January 16, What is theoretically the best classifier

Machine Learning. Theory of Classification and Nonparametric Classifier. Lecture 2, January 16, What is theoretically the best classifier Machine Learning 10-701/15 701/15-781, 781, Spring 2008 Theory of Classification and Nonparametric Classifier Eric Xing Lecture 2, January 16, 2006 Reading: Chap. 2,5 CB and handouts Outline What is theoretically

More information

Reliability and Risk Analysis. Time Series, Types of Trend Functions and Estimates of Trends

Reliability and Risk Analysis. Time Series, Types of Trend Functions and Estimates of Trends Reliability and Risk Analysis Stochastic process The sequence of random variables {Y t, t = 0, ±1, ±2 } is called the stochastic process The mean function of a stochastic process {Y t} is the function

More information

Brownian motion. Samy Tindel. Purdue University. Probability Theory 2 - MA 539

Brownian motion. Samy Tindel. Purdue University. Probability Theory 2 - MA 539 Brownian motion Samy Tindel Purdue University Probability Theory 2 - MA 539 Mostly taken from Brownian Motion and Stochastic Calculus by I. Karatzas and S. Shreve Samy T. Brownian motion Probability Theory

More information

Group Sequential Tests for Delayed Responses. Christopher Jennison. Lisa Hampson. Workshop on Special Topics on Sequential Methodology

Group Sequential Tests for Delayed Responses. Christopher Jennison. Lisa Hampson. Workshop on Special Topics on Sequential Methodology Group Sequential Tests for Delayed Responses Christopher Jennison Department of Mathematical Sciences, University of Bath, UK http://people.bath.ac.uk/mascj Lisa Hampson Department of Mathematics and Statistics,

More information

The Design of Group Sequential Clinical Trials that Test Multiple Endpoints

The Design of Group Sequential Clinical Trials that Test Multiple Endpoints The Design of Group Sequential Clinical Trials that Test Multiple Endpoints Christopher Jennison Department of Mathematical Sciences, University of Bath, UK http://people.bath.ac.uk/mascj Bruce Turnbull

More information

Group Sequential Methods. for Clinical Trials. Christopher Jennison, Dept of Mathematical Sciences, University of Bath, UK

Group Sequential Methods. for Clinical Trials. Christopher Jennison, Dept of Mathematical Sciences, University of Bath, UK Group Sequential Methods for Clinical Trials Christopher Jennison, Dept of Mathematical Sciences, University of Bath, UK http://wwwbathacuk/ mascj University of Malaysia, Kuala Lumpur March 2004 1 Plan

More information

Interim Monitoring of. Clinical Trials. Christopher Jennison, Dept of Mathematical Sciences, University of Bath, UK

Interim Monitoring of. Clinical Trials. Christopher Jennison, Dept of Mathematical Sciences, University of Bath, UK Interim Monitoring of Clinical Trials Christopher Jennison, Dept of Mathematical Sciences, University of Bath, UK http://www.bath.ac.uk/οmascj Magdeburg, 1 October 2003 1 Plan of talk 1. Why sequential

More information

Adaptive Treatment Selection with Survival Endpoints

Adaptive Treatment Selection with Survival Endpoints Adaptive Treatment Selection with Survival Endpoints Gernot Wassmer Institut für Medizinische Statisti, Informati und Epidemiologie Universität zu Köln Joint wor with Marus Roters, Omnicare Clinical Research,

More information

Uniformly Uniformly-ergodic Markov chains and BSDEs

Uniformly Uniformly-ergodic Markov chains and BSDEs Uniformly Uniformly-ergodic Markov chains and BSDEs Samuel N. Cohen Mathematical Institute, University of Oxford (Based on joint work with Ying Hu, Robert Elliott, Lukas Szpruch) Centre Henri Lebesgue,

More information

The strictly 1/2-stable example

The strictly 1/2-stable example The strictly 1/2-stable example 1 Direct approach: building a Lévy pure jump process on R Bert Fristedt provided key mathematical facts for this example. A pure jump Lévy process X is a Lévy process such

More information

Introduction to Empirical Processes and Semiparametric Inference Lecture 02: Overview Continued

Introduction to Empirical Processes and Semiparametric Inference Lecture 02: Overview Continued Introduction to Empirical Processes and Semiparametric Inference Lecture 02: Overview Continued Michael R. Kosorok, Ph.D. Professor and Chair of Biostatistics Professor of Statistics and Operations Research

More information

SAMPLE SIZE RE-ESTIMATION FOR ADAPTIVE SEQUENTIAL DESIGN IN CLINICAL TRIALS

SAMPLE SIZE RE-ESTIMATION FOR ADAPTIVE SEQUENTIAL DESIGN IN CLINICAL TRIALS Journal of Biopharmaceutical Statistics, 18: 1184 1196, 2008 Copyright Taylor & Francis Group, LLC ISSN: 1054-3406 print/1520-5711 online DOI: 10.1080/10543400802369053 SAMPLE SIZE RE-ESTIMATION FOR ADAPTIVE

More information

PhD course in Advanced survival analysis. One-sample tests. Properties. Idea: (ABGK, sect. V.1.1) Counting process N(t)

PhD course in Advanced survival analysis. One-sample tests. Properties. Idea: (ABGK, sect. V.1.1) Counting process N(t) PhD course in Advanced survival analysis. (ABGK, sect. V.1.1) One-sample tests. Counting process N(t) Non-parametric hypothesis tests. Parametric models. Intensity process λ(t) = α(t)y (t) satisfying Aalen

More information

STAT 331. Martingale Central Limit Theorem and Related Results

STAT 331. Martingale Central Limit Theorem and Related Results STAT 331 Martingale Central Limit Theorem and Related Results In this unit we discuss a version of the martingale central limit theorem, which states that under certain conditions, a sum of orthogonal

More information

Introduction to Empirical Processes and Semiparametric Inference Lecture 08: Stochastic Convergence

Introduction to Empirical Processes and Semiparametric Inference Lecture 08: Stochastic Convergence Introduction to Empirical Processes and Semiparametric Inference Lecture 08: Stochastic Convergence Michael R. Kosorok, Ph.D. Professor and Chair of Biostatistics Professor of Statistics and Operations

More information

Model Specification Testing in Nonparametric and Semiparametric Time Series Econometrics. Jiti Gao

Model Specification Testing in Nonparametric and Semiparametric Time Series Econometrics. Jiti Gao Model Specification Testing in Nonparametric and Semiparametric Time Series Econometrics Jiti Gao Department of Statistics School of Mathematics and Statistics The University of Western Australia Crawley

More information

An exponential family of distributions is a parametric statistical model having densities with respect to some positive measure λ of the form.

An exponential family of distributions is a parametric statistical model having densities with respect to some positive measure λ of the form. Stat 8112 Lecture Notes Asymptotics of Exponential Families Charles J. Geyer January 23, 2013 1 Exponential Families An exponential family of distributions is a parametric statistical model having densities

More information

Robust Backtesting Tests for Value-at-Risk Models

Robust Backtesting Tests for Value-at-Risk Models Robust Backtesting Tests for Value-at-Risk Models Jose Olmo City University London (joint work with Juan Carlos Escanciano, Indiana University) Far East and South Asia Meeting of the Econometric Society

More information

Group Sequential Designs: Theory, Computation and Optimisation

Group Sequential Designs: Theory, Computation and Optimisation Group Sequential Designs: Theory, Computation and Optimisation Christopher Jennison Department of Mathematical Sciences, University of Bath, UK http://people.bath.ac.uk/mascj 8th International Conference

More information

ST745: Survival Analysis: Nonparametric methods

ST745: Survival Analysis: Nonparametric methods ST745: Survival Analysis: Nonparametric methods Eric B. Laber Department of Statistics, North Carolina State University February 5, 2015 The KM estimator is used ubiquitously in medical studies to estimate

More information

PROBABILITY: LIMIT THEOREMS II, SPRING HOMEWORK PROBLEMS

PROBABILITY: LIMIT THEOREMS II, SPRING HOMEWORK PROBLEMS PROBABILITY: LIMIT THEOREMS II, SPRING 218. HOMEWORK PROBLEMS PROF. YURI BAKHTIN Instructions. You are allowed to work on solutions in groups, but you are required to write up solutions on your own. Please

More information

Division of Pharmacoepidemiology And Pharmacoeconomics Technical Report Series

Division of Pharmacoepidemiology And Pharmacoeconomics Technical Report Series Division of Pharmacoepidemiology And Pharmacoeconomics Technical Report Series Year: 2013 #006 The Expected Value of Information in Prospective Drug Safety Monitoring Jessica M. Franklin a, Amanda R. Patrick

More information

Malliavin Calculus in Finance

Malliavin Calculus in Finance Malliavin Calculus in Finance Peter K. Friz 1 Greeks and the logarithmic derivative trick Model an underlying assent by a Markov process with values in R m with dynamics described by the SDE dx t = b(x

More information

Sequential Procedure for Testing Hypothesis about Mean of Latent Gaussian Process

Sequential Procedure for Testing Hypothesis about Mean of Latent Gaussian Process Applied Mathematical Sciences, Vol. 4, 2010, no. 62, 3083-3093 Sequential Procedure for Testing Hypothesis about Mean of Latent Gaussian Process Julia Bondarenko Helmut-Schmidt University Hamburg University

More information

INSTABILITY OF RAPIDLY-OSCILLATING PERIODIC SOLUTIONS FOR DISCONTINUOUS DIFFERENTIAL DELAY EQUATIONS

INSTABILITY OF RAPIDLY-OSCILLATING PERIODIC SOLUTIONS FOR DISCONTINUOUS DIFFERENTIAL DELAY EQUATIONS Differential and Integral Equations Volume 15, Number 1, January 2002, Pages 53 90 INSTABILITY OF RAPIDLY-OSCILLATING PERIODIC SOLUTIONS FOR DISCONTINUOUS DIFFERENTIAL DELAY EQUATIONS Marianne Akian INRIA,

More information

Adaptive Designs: Why, How and When?

Adaptive Designs: Why, How and When? Adaptive Designs: Why, How and When? Christopher Jennison Department of Mathematical Sciences, University of Bath, UK http://people.bath.ac.uk/mascj ISBS Conference Shanghai, July 2008 1 Adaptive designs:

More information

Reliability Theory of Dynamic Loaded Structures (cont.) Calculation of Out-Crossing Frequencies Approximations to the Failure Probability.

Reliability Theory of Dynamic Loaded Structures (cont.) Calculation of Out-Crossing Frequencies Approximations to the Failure Probability. Outline of Reliability Theory of Dynamic Loaded Structures (cont.) Calculation of Out-Crossing Frequencies Approximations to the Failure Probability. Poisson Approximation. Upper Bound Solution. Approximation

More information

Reflected Brownian Motion

Reflected Brownian Motion Chapter 6 Reflected Brownian Motion Often we encounter Diffusions in regions with boundary. If the process can reach the boundary from the interior in finite time with positive probability we need to decide

More information

A Type of Sample Size Planning for Mean Comparison in Clinical Trials

A Type of Sample Size Planning for Mean Comparison in Clinical Trials Journal of Data Science 13(2015), 115-126 A Type of Sample Size Planning for Mean Comparison in Clinical Trials Junfeng Liu 1 and Dipak K. Dey 2 1 GCE Solutions, Inc. 2 Department of Statistics, University

More information

Minicourse on: Markov Chain Monte Carlo: Simulation Techniques in Statistics

Minicourse on: Markov Chain Monte Carlo: Simulation Techniques in Statistics Minicourse on: Markov Chain Monte Carlo: Simulation Techniques in Statistics Eric Slud, Statistics Program Lecture 1: Metropolis-Hastings Algorithm, plus background in Simulation and Markov Chains. Lecture

More information

University of Warwick, EC9A0 Maths for Economists Lecture Notes 10: Dynamic Programming

University of Warwick, EC9A0 Maths for Economists Lecture Notes 10: Dynamic Programming University of Warwick, EC9A0 Maths for Economists 1 of 63 University of Warwick, EC9A0 Maths for Economists Lecture Notes 10: Dynamic Programming Peter J. Hammond Autumn 2013, revised 2014 University of

More information

Statistical Learning Theory and the C-Loss cost function

Statistical Learning Theory and the C-Loss cost function Statistical Learning Theory and the C-Loss cost function Jose Principe, Ph.D. Distinguished Professor ECE, BME Computational NeuroEngineering Laboratory and principe@cnel.ufl.edu Statistical Learning Theory

More information

ELEMENTS OF PROBABILITY THEORY

ELEMENTS OF PROBABILITY THEORY ELEMENTS OF PROBABILITY THEORY Elements of Probability Theory A collection of subsets of a set Ω is called a σ algebra if it contains Ω and is closed under the operations of taking complements and countable

More information

Lecture Notes 10: Dynamic Programming

Lecture Notes 10: Dynamic Programming University of Warwick, EC9A0 Maths for Economists Peter J. Hammond 1 of 81 Lecture Notes 10: Dynamic Programming Peter J. Hammond 2018 September 28th University of Warwick, EC9A0 Maths for Economists Peter

More information

University of California, Berkeley

University of California, Berkeley University of California, Berkeley U.C. Berkeley Division of Biostatistics Working Paper Series Year 24 Paper 153 A Note on Empirical Likelihood Inference of Residual Life Regression Ying Qing Chen Yichuan

More information

On detection of unit roots generalizing the classic Dickey-Fuller approach

On detection of unit roots generalizing the classic Dickey-Fuller approach On detection of unit roots generalizing the classic Dickey-Fuller approach A. Steland Ruhr-Universität Bochum Fakultät für Mathematik Building NA 3/71 D-4478 Bochum, Germany February 18, 25 1 Abstract

More information

Adaptive Extensions of a Two-Stage Group Sequential Procedure for Testing a Primary and a Secondary Endpoint (II): Sample Size Re-estimation

Adaptive Extensions of a Two-Stage Group Sequential Procedure for Testing a Primary and a Secondary Endpoint (II): Sample Size Re-estimation Research Article Received XXXX (www.interscience.wiley.com) DOI: 10.100/sim.0000 Adaptive Extensions of a Two-Stage Group Sequential Procedure for Testing a Primary and a Secondary Endpoint (II): Sample

More information

Sequential Plans and Risk Evaluation

Sequential Plans and Risk Evaluation C. Schmegner and M. Baron. Sequential Plans and Risk Evaluation. Sequential Analysis, 26(4), 335 354, 2007. Sequential Plans and Risk Evaluation Claudia Schmegner Department of Mathematical Sciences, DePaul

More information

Hypothesis testing for Stochastic PDEs. Igor Cialenco

Hypothesis testing for Stochastic PDEs. Igor Cialenco Hypothesis testing for Stochastic PDEs Igor Cialenco Department of Applied Mathematics Illinois Institute of Technology igor@math.iit.edu Joint work with Liaosha Xu Research partially funded by NSF grants

More information

Monitoring clinical trial outcomes with delayed response: Incorporating pipeline data in group sequential and adaptive designs. Christopher Jennison

Monitoring clinical trial outcomes with delayed response: Incorporating pipeline data in group sequential and adaptive designs. Christopher Jennison Monitoring clinical trial outcomes with delayed response: Incorporating pipeline data in group sequential and adaptive designs Christopher Jennison Department of Mathematical Sciences, University of Bath,

More information

PRINCIPLES OF OPTIMAL SEQUENTIAL PLANNING

PRINCIPLES OF OPTIMAL SEQUENTIAL PLANNING C. Schmegner and M. Baron. Principles of optimal sequential planning. Sequential Analysis, 23(1), 11 32, 2004. Abraham Wald Memorial Issue PRINCIPLES OF OPTIMAL SEQUENTIAL PLANNING Claudia Schmegner 1

More information

Statistical Inference and Methods

Statistical Inference and Methods Department of Mathematics Imperial College London d.stephens@imperial.ac.uk http://stats.ma.ic.ac.uk/ das01/ 31st January 2006 Part VI Session 6: Filtering and Time to Event Data Session 6: Filtering and

More information

LAN property for ergodic jump-diffusion processes with discrete observations

LAN property for ergodic jump-diffusion processes with discrete observations LAN property for ergodic jump-diffusion processes with discrete observations Eulalia Nualart (Universitat Pompeu Fabra, Barcelona) joint work with Arturo Kohatsu-Higa (Ritsumeikan University, Japan) &

More information

Poisson random measure: motivation

Poisson random measure: motivation : motivation The Lévy measure provides the expected number of jumps by time unit, i.e. in a time interval of the form: [t, t + 1], and of a certain size Example: ν([1, )) is the expected number of jumps

More information

Control of Directional Errors in Fixed Sequence Multiple Testing

Control of Directional Errors in Fixed Sequence Multiple Testing Control of Directional Errors in Fixed Sequence Multiple Testing Anjana Grandhi Department of Mathematical Sciences New Jersey Institute of Technology Newark, NJ 07102-1982 Wenge Guo Department of Mathematical

More information

Mean-Field optimization problems and non-anticipative optimal transport. Beatrice Acciaio

Mean-Field optimization problems and non-anticipative optimal transport. Beatrice Acciaio Mean-Field optimization problems and non-anticipative optimal transport Beatrice Acciaio London School of Economics based on ongoing projects with J. Backhoff, R. Carmona and P. Wang Robust Methods in

More information

Stochastic Volatility and Correction to the Heat Equation

Stochastic Volatility and Correction to the Heat Equation Stochastic Volatility and Correction to the Heat Equation Jean-Pierre Fouque, George Papanicolaou and Ronnie Sircar Abstract. From a probabilist s point of view the Twentieth Century has been a century

More information

Survival Analysis for Case-Cohort Studies

Survival Analysis for Case-Cohort Studies Survival Analysis for ase-ohort Studies Petr Klášterecký Dept. of Probability and Mathematical Statistics, Faculty of Mathematics and Physics, harles University, Prague, zech Republic e-mail: petr.klasterecky@matfyz.cz

More information

Econ 2148, fall 2017 Gaussian process priors, reproducing kernel Hilbert spaces, and Splines

Econ 2148, fall 2017 Gaussian process priors, reproducing kernel Hilbert spaces, and Splines Econ 2148, fall 2017 Gaussian process priors, reproducing kernel Hilbert spaces, and Splines Maximilian Kasy Department of Economics, Harvard University 1 / 37 Agenda 6 equivalent representations of the

More information

6 Basic Convergence Results for RL Algorithms

6 Basic Convergence Results for RL Algorithms Learning in Complex Systems Spring 2011 Lecture Notes Nahum Shimkin 6 Basic Convergence Results for RL Algorithms We establish here some asymptotic convergence results for the basic RL algorithms, by showing

More information

Group Sequential Methods. for Clinical Trials. Christopher Jennison, Dept of Mathematical Sciences, University of Bath, UK

Group Sequential Methods. for Clinical Trials. Christopher Jennison, Dept of Mathematical Sciences, University of Bath, UK Group Sequential Methods for Clinical Trials Christopher Jennison, Dept of Mathematical Sciences, University of Bath, UK http://www.bath.ac.uk/οmascj PSI, London, 22 October 2003 1 Plan of talk 1. Why

More information

Economics 2010c: Lectures 9-10 Bellman Equation in Continuous Time

Economics 2010c: Lectures 9-10 Bellman Equation in Continuous Time Economics 2010c: Lectures 9-10 Bellman Equation in Continuous Time David Laibson 9/30/2014 Outline Lectures 9-10: 9.1 Continuous-time Bellman Equation 9.2 Application: Merton s Problem 9.3 Application:

More information

More Empirical Process Theory

More Empirical Process Theory More Empirical Process heory 4.384 ime Series Analysis, Fall 2008 Recitation by Paul Schrimpf Supplementary to lectures given by Anna Mikusheva October 24, 2008 Recitation 8 More Empirical Process heory

More information

Chapter 2: Fundamentals of Statistics Lecture 15: Models and statistics

Chapter 2: Fundamentals of Statistics Lecture 15: Models and statistics Chapter 2: Fundamentals of Statistics Lecture 15: Models and statistics Data from one or a series of random experiments are collected. Planning experiments and collecting data (not discussed here). Analysis:

More information

Group Sequential Tests for Delayed Responses

Group Sequential Tests for Delayed Responses Group Sequential Tests for Delayed Responses Lisa Hampson Department of Mathematics and Statistics, Lancaster University, UK Chris Jennison Department of Mathematical Sciences, University of Bath, UK Read

More information

Bios 6649: Clinical Trials - Statistical Design and Monitoring

Bios 6649: Clinical Trials - Statistical Design and Monitoring Bios 6649: Clinical Trials - Statistical Design and Monitoring Spring Semester 2015 John M. Kittelson Department of Biostatistics & Informatics Colorado School of Public Health University of Colorado Denver

More information

A Barrier Version of the Russian Option

A Barrier Version of the Russian Option A Barrier Version of the Russian Option L. A. Shepp, A. N. Shiryaev, A. Sulem Rutgers University; shepp@stat.rutgers.edu Steklov Mathematical Institute; shiryaev@mi.ras.ru INRIA- Rocquencourt; agnes.sulem@inria.fr

More information

Simultaneous Group Sequential Analysis of Rank-Based and Weighted Kaplan Meier Tests for Paired Censored Survival Data

Simultaneous Group Sequential Analysis of Rank-Based and Weighted Kaplan Meier Tests for Paired Censored Survival Data Biometrics 61, 715 72 September 25 DOI: 11111/j1541-4225337x Simultaneous Group Sequential Analysis of Rank-Based and Weighted Kaplan Meier Tests for Paired Censored Survival Data Adin-Cristian Andrei

More information

Mean-Field optimization problems and non-anticipative optimal transport. Beatrice Acciaio

Mean-Field optimization problems and non-anticipative optimal transport. Beatrice Acciaio Mean-Field optimization problems and non-anticipative optimal transport Beatrice Acciaio London School of Economics based on ongoing projects with J. Backhoff, R. Carmona and P. Wang Thera Stochastics

More information

1 Probability Model. 1.1 Types of models to be discussed in the course

1 Probability Model. 1.1 Types of models to be discussed in the course Sufficiency January 18, 016 Debdeep Pati 1 Probability Model Model: A family of distributions P θ : θ Θ}. P θ (B) is the probability of the event B when the parameter takes the value θ. P θ is described

More information

Applications of Ito s Formula

Applications of Ito s Formula CHAPTER 4 Applications of Ito s Formula In this chapter, we discuss several basic theorems in stochastic analysis. Their proofs are good examples of applications of Itô s formula. 1. Lévy s martingale

More information

Confidence Intervals of Prescribed Precision Summary

Confidence Intervals of Prescribed Precision Summary Confidence Intervals of Prescribed Precision Summary Charles Stein showed in 1945 that by using a two stage sequential procedure one could give a confidence interval for the mean of a normal distribution

More information

The multidimensional Ito Integral and the multidimensional Ito Formula. Eric Mu ller June 1, 2015 Seminar on Stochastic Geometry and its applications

The multidimensional Ito Integral and the multidimensional Ito Formula. Eric Mu ller June 1, 2015 Seminar on Stochastic Geometry and its applications The multidimensional Ito Integral and the multidimensional Ito Formula Eric Mu ller June 1, 215 Seminar on Stochastic Geometry and its applications page 2 Seminar on Stochastic Geometry and its applications

More information

Information and Credit Risk

Information and Credit Risk Information and Credit Risk M. L. Bedini Université de Bretagne Occidentale, Brest - Friedrich Schiller Universität, Jena Jena, March 2011 M. L. Bedini (Université de Bretagne Occidentale, Brest Information

More information

Closest Moment Estimation under General Conditions

Closest Moment Estimation under General Conditions Closest Moment Estimation under General Conditions Chirok Han Victoria University of Wellington New Zealand Robert de Jong Ohio State University U.S.A October, 2003 Abstract This paper considers Closest

More information

Math 680 Fall A Quantitative Prime Number Theorem I: Zero-Free Regions

Math 680 Fall A Quantitative Prime Number Theorem I: Zero-Free Regions Math 68 Fall 4 A Quantitative Prime Number Theorem I: Zero-Free Regions Ultimately, our goal is to prove the following strengthening of the prime number theorem Theorem Improved Prime Number Theorem: There

More information

Hypothesis Testing Based on the Maximum of Two Statistics from Weighted and Unweighted Estimating Equations

Hypothesis Testing Based on the Maximum of Two Statistics from Weighted and Unweighted Estimating Equations Hypothesis Testing Based on the Maximum of Two Statistics from Weighted and Unweighted Estimating Equations Takeshi Emura and Hisayuki Tsukuma Abstract For testing the regression parameter in multivariate

More information

Lecture 17 Brownian motion as a Markov process

Lecture 17 Brownian motion as a Markov process Lecture 17: Brownian motion as a Markov process 1 of 14 Course: Theory of Probability II Term: Spring 2015 Instructor: Gordan Zitkovic Lecture 17 Brownian motion as a Markov process Brownian motion is

More information

The information complexity of sequential resource allocation

The information complexity of sequential resource allocation The information complexity of sequential resource allocation Emilie Kaufmann, joint work with Olivier Cappé, Aurélien Garivier and Shivaram Kalyanakrishan SMILE Seminar, ENS, June 8th, 205 Sequential allocation

More information

Group sequential designs with negative binomial data

Group sequential designs with negative binomial data Group sequential designs with negative binomial data Ekkehard Glimm 1 Tobias Mütze 2,3 1 Statistical Methodology, Novartis, Basel, Switzerland 2 Department of Medical Statistics, University Medical Center

More information

Lecture 21. Hypothesis Testing II

Lecture 21. Hypothesis Testing II Lecture 21. Hypothesis Testing II December 7, 2011 In the previous lecture, we dened a few key concepts of hypothesis testing and introduced the framework for parametric hypothesis testing. In the parametric

More information

Global Maxwellians over All Space and Their Relation to Conserved Quantites of Classical Kinetic Equations

Global Maxwellians over All Space and Their Relation to Conserved Quantites of Classical Kinetic Equations Global Maxwellians over All Space and Their Relation to Conserved Quantites of Classical Kinetic Equations C. David Levermore Department of Mathematics and Institute for Physical Science and Technology

More information

Solution of Stochastic Optimal Control Problems and Financial Applications

Solution of Stochastic Optimal Control Problems and Financial Applications Journal of Mathematical Extension Vol. 11, No. 4, (2017), 27-44 ISSN: 1735-8299 URL: http://www.ijmex.com Solution of Stochastic Optimal Control Problems and Financial Applications 2 Mat B. Kafash 1 Faculty

More information

ROI ANALYSIS OF PHARMAFMRI DATA:

ROI ANALYSIS OF PHARMAFMRI DATA: ROI ANALYSIS OF PHARMAFMRI DATA: AN ADAPTIVE APPROACH FOR GLOBAL TESTING Giorgos Minas, John A.D. Aston, Thomas E. Nichols and Nigel Stallard Department of Statistics and Warwick Centre of Analytical Sciences,

More information

Asymptotic statistics using the Functional Delta Method

Asymptotic statistics using the Functional Delta Method Quantiles, Order Statistics and L-Statsitics TU Kaiserslautern 15. Februar 2015 Motivation Functional The delta method introduced in chapter 3 is an useful technique to turn the weak convergence of random

More information

Sequential Decisions

Sequential Decisions Sequential Decisions A Basic Theorem of (Bayesian) Expected Utility Theory: If you can postpone a terminal decision in order to observe, cost free, an experiment whose outcome might change your terminal

More information

SEQUENTIAL MODELS FOR CLINICAL TRIALS. patients to be treated. anticipated that a horizon of N patients will have to be treated by one drug or

SEQUENTIAL MODELS FOR CLINICAL TRIALS. patients to be treated. anticipated that a horizon of N patients will have to be treated by one drug or 1. Introduction SEQUENTIAL MODELS FOR CLINICAL TRIALS HERMAN CHERNOFF STANFORD UNIVERSITY This paper presents a model, variations of which have been considered by Anscombe [1] and Colton [4] and others,

More information

Lecture 12. F o s, (1.1) F t := s>t

Lecture 12. F o s, (1.1) F t := s>t Lecture 12 1 Brownian motion: the Markov property Let C := C(0, ), R) be the space of continuous functions mapping from 0, ) to R, in which a Brownian motion (B t ) t 0 almost surely takes its value. Let

More information

Regularity of the density for the stochastic heat equation

Regularity of the density for the stochastic heat equation Regularity of the density for the stochastic heat equation Carl Mueller 1 Department of Mathematics University of Rochester Rochester, NY 15627 USA email: cmlr@math.rochester.edu David Nualart 2 Department

More information

Chapter 9: Hypothesis Testing Sections

Chapter 9: Hypothesis Testing Sections Chapter 9: Hypothesis Testing Sections 9.1 Problems of Testing Hypotheses 9.2 Testing Simple Hypotheses 9.3 Uniformly Most Powerful Tests Skip: 9.4 Two-Sided Alternatives 9.6 Comparing the Means of Two

More information