Kostas Triantafyllopoulos University of Sheffield. Motivation / algorithmic pairs trading. Model set-up Detection of local mean-reversion
|
|
- Todd Rudolph Griffith
- 5 years ago
- Views:
Transcription
1 Detecting Mean Reverted Patterns in Statistical Arbitrage Outline Kostas Triantafyllopoulos University of Sheffield Motivation / algorithmic pairs trading Model set-up Detection of local mean-reversion Adaptive estimation 1. RLS with gradient variable forgetting factor 2. RLS with Gauss-Newton variable forgetting factor 3. RLS with beta-bernoulli forgetting factor Trading strategy Pepsi and Coca Cola example
2 Introduction Statistical arbitrage. Algorithmic pairs trading market neutral trading. Buy low, sell high. Two assets with prices p At and p Bt At t: if p At < p Bt, buy low (A) and sell high (B). At t + 1: if pat > p Bt, buy low (B) and sell high (A)... In the long run, mean reversion of spread y t = p At p Bt. If y t goes up, y t will go down at t + 1. Take advantage of relative mispricings of A and B.
3 Introduction Share price stream Share price A,B Time Spread stream A B Time
4 Pepsi - Coca Cola date stream Share Prices (in USD) Mean of Spread = Pepsi Coca Cola Trading day
5 Model set-up (Elliott et al, 2005). y t is a noisy version of a mean-reverted process. y t = x t + ε t x t = α + βx t 1 + ζ t (Triantafyllopoulos and Montana, 2011). y t = α t + β t y t 1 + ϵ t = F T t θ t + ϵ t, θ t = Φθ t 1 + ω t, [ ] 1 F t = and θ y t = t 1 [ αt β t ]
6 Detecting mean reversion Define D t = (y 1,..., y t ) a data stream sample. (Elliott et al, 2005). If β < 1, then y t is mean-reverted. With D t, get online estimates ˆα and ˆβ. If ˆβ < 1, then mean-reversion. (Triantafyllopoulos and Montana, 2011). Under some assumptions y t is mean-reverted if β t < 1, for all t. We consider mean-reversion in segments or locally. Test again ˆβ t < 1. So we need online estimates ˆβ t.
7 Recursive least squares (RLS) (Haykin, 2001). Recursive Least Squares (RLS), if Φ = I. Find θ that minimizes the cost function t 1 λ j (y t j Ft jθ) T 2 j=0 0 < λ < 1 a forgetting factor. Past data are forgotten at a rate t 1 λ j = 1 + λ + + λ t λ j=0 [memory of the system] With λ = 1, ordinary regression, memory is With λ < 1, RLS, memory (1 λ) 1.
8 SS-RLS (state-space RLS) Variable forgetting factor (Haykin, 2001, Malik, 2006). Steepest descent: λ t λ t 1 = c(t) λ t = [λ t 1 + a λ (t)] λ+ λ, 1. λ (t) e t F T t Φψ t 1 2. ψ t = m t / λ = (I K t F T t )Φψ t 1 + S t F t e t 3. S t = P t / λ = λ 1 t P t + λ 1 t K t Kt T +λ 1 t (I K t Ft T )ΦS t 1 Φ T (I F t Kt T ). Need starting values m 1, P 1, ψ 1, S 1, λ 1.
9 GN-RLS (Gauss-Newton RLS) Song et al (2000) gave an approximate GN algorithm. λ t = [ λ t 1 + a ] λ+ λ(t) 2 λ (t), λ Here 2 λ (t) (F T t Φψ t 1 ) 2 e t F T t Φη t η t = ψ t 1 / λ = (I K t F T t )Φη t 1 + L t F t e t 2S t F t F T t Φψ t 1 where and L t = λ 1 t (I K t Ft T )ΦL t 1 Φ T (I F t Kt T ) +λ 2 t P t (I F t Kt T ) λ 1 t S t + M t + Mt T λ 2 t (I K t Ft T )ΦS t 1 Φ T (I F t Kt T ) M t = λ 1 t S t F t F T t {P t ΦS t 1 Φ T (I F t K T t )}.
10 GN-RLS cont. GN-RLS creates too abrupt jumps in λ t / too sensitive to changes. When smooth signal, we want SS-RLS and when noisy we want GN-RLS. We use λ t = [λ t 1 + a λ (t)] λ+ λ, if e2 t k [ λ t 1 + a λ(t) 2 λ (t) ] λ+ λ, if e2 t > k
11 BB-RLS (beta Bernoulli RLS) In the above we ask λ λ t λ +. If et 2 small, λ t λ + (smoothness) If et 2 large, λ t λ (adaptiveness) We set λ t = πλ + + (1 π)λ, π = Pr(et 2 k t ) Two events small prediction error / large prediction error { 1, if e 2 x t = t k t, with probability π 0, if et 2 > k t, with probability 1 π
12 Observation model x t Bernoulli(π). Prior for π is beta, π B(c 1, c 2 ) p(x t π) = π xt (1 x t ) 1 xt p(π) π c 1 1 (1 π) c 2 1 (bernoulli model) (prior beta model) p(π x t ) p(x t π)p(π) π c 1+x t 1 (1 π) c 2+1 x t 1 So π x t B(c 1 + x t, c 2 x t + 1). Sequentially: π x 1,..., x t π D t 1 Be(c 1t, c 2t ), where c 1t = c 1,t 1 + x t and c 2t = c 2,t 1 x t + 1. ˆπ t = E(π D t 1 ) = c 1t (c 1t + c 2t ) 1 ˆλ t = E(λ t D t 1 ) = ˆπ t λ + + (1 ˆπ t )λ From Pr(e 2 t k t ) = π, we have k t = q t F 1 χ 2 (π). We use k t q t F 1 χ 2 (ˆπ t 1 ).
13 Key points of λ t λ t is stochastic. We can derive its distribution p(λ t D t 1 ) = c(λ t λ ) c 1t 1 (λ + λ t ) c 2t 1 We can evaluate the mode and the variance of λ t. We can show ˆλ t xλ + + (1 x)λ If for many points e 2 t < k t, followed by few large e 2 t > k t, then locally ˆλ t does not work well.
14 Solution: intervention Example. 1. Set λ = 0.8, λ + = e 2 t k t (t = 1,..., 95) 3. e 2 t > k t (t = 96, 97, 98, 99, 100) 4. ˆλ t = (closer to λ + = 0.99, than to λ = 0.8). If change in x t (from 0 to 1 or from 1 to 0), reset the priors c 1,t 1 and c 2,t 1 to the initial values (c 1,1 = c 2,1 = 0.5). ˆπ t = c 1t c 1t + c 2t = c 1,1 + x t c 1,1 + c 2,1 + 1 = 1 + 2x t 4 ˆλ t = { 0.75λ λ, if x t = λ λ, if x t = 0 In the example, ˆλ 96 = =
15 Simulated streams Simulated Spread Time Prediction of B t SS RLS GN RLS BB RLS Trading day
16 Trading strategy Spread y t = p A,t p B,t. If y t not mean-reverted do nothing. It is not predictable. If y t < y t+1, p A,t+1 > p A,t or p B,t+1 < p B,t. Buy A / sell B. If y t > y t+1, p A,t+1 < p A,t or p B,t+1 > p B,t. Buy B / sell A. If y t y t 1, do nothing. At time t, we don t know y t+1, so we predict it and use ŷ t+1.
17 Trading strategy Trading strategy p(a) > p(b) Tradable time spread y t 1 t Non tradable time p(a) < p(b) t+1
18 With observed spread y t at t: Close the position of t 1 (if opened). If ˆβ t+1 < 0.99, declare y t+1 as mean-reverted. If ŷ t+1 h > y t, buy A / short-sell B. If ŷ t+1 + h < y t, buy B / short-sell A. Example: At t: y t = 10 and we project ŷ t+1 = 11. y t+1 can be 12 or 9 (rules change if we apply y t <> ŷ t+1. y t+1 = 9 can give loss, if we adopt y t < ŷ t+1 rule (10 < 11). Take h = 10% of ŷ t+1 = 1.1, and ŷ t+1 h = 9.9 < 10 = y t, we do not open a position buy A and short sell B. As ŷ t+1 + h = 11.1 > 10 = y t, we do not open a position short sell A and buy B.
19 Pepsi - Coca Cola date stream Share Prices (in USD) Mean of Spread = Pepsi Coca Cola Trading day
20 Detection of mean reverted patterns Prediction of B t BB RLS SS RLS GN RLS Trading day
21 MSE over time Mean square error of the three algorithms SS RLS GN RLS BB RLS Trading day
22 Trading performance Mean 1% 3% 5% SS-RLS GN-RLS BB-RLS STD 1% 3% 5% SS-RLS GN-RLS BB-RLS FB 1% 3% 5% SS-RLS GN-RLS BB-RLS h
23 Trading performance Cumulative profit BB RLS GN RLS SS RLS Trading day
24 Closing remarks Algorithmic pairs trading / statistical arbitrage require online machine learning methods. Pattern recognition methods for mean reversion / segments of stationarity. We develop variable forgetting factors for online learning. Other methods include sequential Monte Carlo. Need to take into account the shape of the distribution of the data stream. Larger data streams / complex portfolios / many pairs to consider simultaneously. Trading strategy can be improved.
25 References Elliott, R., Van Der Hoek, J., and Malcolm, W. (2005). Pairs trading. Quantitative Finance, 5: Haykin, S. (2001). Adaptive Filter Theory. Prentice Hall, 4th edition. Malik, M. B. (2006). State-space recursive least-squares with adaptive memory. Signal Processing, 86: Song, S., Lim, J.-S., Baek, S., and Sung, K.-M. (2000). Gauss-Newton variable forgetting factor recursive least squares for time varying parameter tracking. Electronics Letters, 36: Triantafyllopoulos, K. and Montana, G. (2011). Dynamic modeling of mean-reverting spreads for statistical arbitrage. Computational Management Science, 8:23-49.
Comparative Performance Analysis of Three Algorithms for Principal Component Analysis
84 R. LANDQVIST, A. MOHAMMED, COMPARATIVE PERFORMANCE ANALYSIS OF THR ALGORITHMS Comparative Performance Analysis of Three Algorithms for Principal Component Analysis Ronnie LANDQVIST, Abbas MOHAMMED Dept.
More informationCondensed Table of Contents for Introduction to Stochastic Search and Optimization: Estimation, Simulation, and Control by J. C.
Condensed Table of Contents for Introduction to Stochastic Search and Optimization: Estimation, Simulation, and Control by J. C. Spall John Wiley and Sons, Inc., 2003 Preface... xiii 1. Stochastic Search
More informationThe Slow Convergence of OLS Estimators of α, β and Portfolio. β and Portfolio Weights under Long Memory Stochastic Volatility
The Slow Convergence of OLS Estimators of α, β and Portfolio Weights under Long Memory Stochastic Volatility New York University Stern School of Business June 21, 2018 Introduction Bivariate long memory
More informationErrata for Campbell, Financial Decisions and Markets, 01/02/2019.
Errata for Campbell, Financial Decisions and Markets, 01/02/2019. Page xi, section title for Section 11.4.3 should be Endogenous Margin Requirements. Page 20, equation 1.49), expectations operator E should
More informationRecursive Generalized Eigendecomposition for Independent Component Analysis
Recursive Generalized Eigendecomposition for Independent Component Analysis Umut Ozertem 1, Deniz Erdogmus 1,, ian Lan 1 CSEE Department, OGI, Oregon Health & Science University, Portland, OR, USA. {ozertemu,deniz}@csee.ogi.edu
More informationOptimal combination forecasts for hierarchical time series
Optimal combination forecasts for hierarchical time series Rob J. Hyndman Roman A. Ahmed Department of Econometrics and Business Statistics Outline 1 Review of hierarchical forecasting 2 3 Simulation study
More informationRecent Advances in SPSA at the Extremes: Adaptive Methods for Smooth Problems and Discrete Methods for Non-Smooth Problems
Recent Advances in SPSA at the Extremes: Adaptive Methods for Smooth Problems and Discrete Methods for Non-Smooth Problems SGM2014: Stochastic Gradient Methods IPAM, February 24 28, 2014 James C. Spall
More informationMachine Learning CS 4900/5900. Lecture 03. Razvan C. Bunescu School of Electrical Engineering and Computer Science
Machine Learning CS 4900/5900 Razvan C. Bunescu School of Electrical Engineering and Computer Science bunescu@ohio.edu Machine Learning is Optimization Parametric ML involves minimizing an objective function
More informationLecture 2: From Linear Regression to Kalman Filter and Beyond
Lecture 2: From Linear Regression to Kalman Filter and Beyond January 18, 2017 Contents 1 Batch and Recursive Estimation 2 Towards Bayesian Filtering 3 Kalman Filter and Bayesian Filtering and Smoothing
More informationLinear Regression (continued)
Linear Regression (continued) Professor Ameet Talwalkar Professor Ameet Talwalkar CS260 Machine Learning Algorithms February 6, 2017 1 / 39 Outline 1 Administration 2 Review of last lecture 3 Linear regression
More informationAssessing others rationality in real time
Assessing others rationality in real time Gaetano GABALLO Ph.D.candidate University of Siena, Italy June 4, 2009 Gaetano GABALLO Ph.D.candidate University of Siena, Italy Assessing () others rationality
More informationDynamic Risk Measures and Nonlinear Expectations with Markov Chain noise
Dynamic Risk Measures and Nonlinear Expectations with Markov Chain noise Robert J. Elliott 1 Samuel N. Cohen 2 1 Department of Commerce, University of South Australia 2 Mathematical Insitute, University
More informationLogistic Regression. Stochastic Gradient Descent
Tutorial 8 CPSC 340 Logistic Regression Stochastic Gradient Descent Logistic Regression Model A discriminative probabilistic model for classification e.g. spam filtering Let x R d be input and y { 1, 1}
More informationLinear Models for Regression
Linear Models for Regression Seungjin Choi Department of Computer Science and Engineering Pohang University of Science and Technology 77 Cheongam-ro, Nam-gu, Pohang 37673, Korea seungjin@postech.ac.kr
More informationHigh-dimensional Problems in Finance and Economics. Thomas M. Mertens
High-dimensional Problems in Finance and Economics Thomas M. Mertens NYU Stern Risk Economics Lab April 17, 2012 1 / 78 Motivation Many problems in finance and economics are high dimensional. Dynamic Optimization:
More informationFitting a regression model
Fitting a regression model We wish to fit a simple linear regression model: y = β 0 + β 1 x + ɛ. Fitting a model means obtaining estimators for the unknown population parameters β 0 and β 1 (and also for
More informationMonitoring Forecasting Performance
Monitoring Forecasting Performance Identifying when and why return prediction models work Allan Timmermann and Yinchu Zhu University of California, San Diego June 21, 2015 Outline Testing for time-varying
More informationMAS113 Introduction to Probability and Statistics
MAS113 Introduction to Probability and Statistics School of Mathematics and Statistics, University of Sheffield 2018 19 Identically distributed Suppose we have n random variables X 1, X 2,..., X n. Identically
More informationLecture 10. Neural networks and optimization. Machine Learning and Data Mining November Nando de Freitas UBC. Nonlinear Supervised Learning
Lecture 0 Neural networks and optimization Machine Learning and Data Mining November 2009 UBC Gradient Searching for a good solution can be interpreted as looking for a minimum of some error (loss) function
More informationMotivation Subgradient Method Stochastic Subgradient Method. Convex Optimization. Lecture 15 - Gradient Descent in Machine Learning
Convex Optimization Lecture 15 - Gradient Descent in Machine Learning Instructor: Yuanzhang Xiao University of Hawaii at Manoa Fall 2017 1 / 21 Today s Lecture 1 Motivation 2 Subgradient Method 3 Stochastic
More informationarxiv: v1 [q-fin.st] 25 Sep 2007
Flexible least squares for temporal data mining arxiv:0709.3884v1 [q-fin.st] 25 Sep 2007 and statistical arbitrage Giovanni Montana, Kostas Triantafyllopoulos, Theodoros Tsagaris December 3, 2008 Abstract
More informationASSET PRICING MODELS
ASSE PRICING MODELS [1] CAPM (1) Some notation: R it = (gross) return on asset i at time t. R mt = (gross) return on the market portfolio at time t. R ft = return on risk-free asset at time t. X it = R
More informationIntroduction to Algorithmic Trading Strategies Lecture 3
Introduction to Algorithmic Trading Strategies Lecture 3 Pairs Trading by Cointegration Haksun Li haksun.li@numericalmethod.com www.numericalmethod.com Outline Distance method Cointegration Stationarity
More informationSome Notes from the Book: Pairs Trading: Quantitative Methods and Analysis by Ganapathy Vidyamurthy
1 Some Notes from the Book: Pairs Trading: Quantitative Methods and Analysis by Ganapathy Vidyamurthy John L. Weatherwax Sept 30, 2004 wax@alum.mit.edu 2 Chapter 1 (Introduction) Notes on a Market Neutral
More informationIntroduction General Framework Toy models Discrete Markov model Data Analysis Conclusion. The Micro-Price. Sasha Stoikov. Cornell University
The Micro-Price Sasha Stoikov Cornell University Jim Gatheral @ NYU High frequency traders (HFT) HFTs are good: Optimal order splitting Pairs trading / statistical arbitrage Market making / liquidity provision
More informationSwitching Regime Estimation
Switching Regime Estimation Series de Tiempo BIrkbeck March 2013 Martin Sola (FE) Markov Switching models 01/13 1 / 52 The economy (the time series) often behaves very different in periods such as booms
More informationClassification: Logistic Regression from Data
Classification: Logistic Regression from Data Machine Learning: Jordan Boyd-Graber University of Colorado Boulder LECTURE 3 Slides adapted from Emily Fox Machine Learning: Jordan Boyd-Graber Boulder Classification:
More informationFinancial Econometrics
Financial Econometrics Nonlinear time series analysis Gerald P. Dwyer Trinity College, Dublin January 2016 Outline 1 Nonlinearity Does nonlinearity matter? Nonlinear models Tests for nonlinearity Forecasting
More informationarxiv: v1 [q-fin.st] 12 Aug 2008
Dynamic modeling of mean-reverting spreads for statistical arxiv:0808.1710v1 [q-fin.st] 12 Aug 2008 arbitrage K. Triantafyllopoulos G. Montana May 29, 2009 Abstract Statistical arbitrage strategies, such
More informationLecture 13. Simple Linear Regression
1 / 27 Lecture 13 Simple Linear Regression October 28, 2010 2 / 27 Lesson Plan 1. Ordinary Least Squares 2. Interpretation 3 / 27 Motivation Suppose we want to approximate the value of Y with a linear
More informationThe Extended Kalman Filter is a Natural Gradient Descent in Trajectory Space
The Extended Kalman Filter is a Natural Gradient Descent in Trajectory Space Yann Ollivier Abstract The extended Kalman filter is perhaps the most standard tool to estimate in real time the state of a
More informationNonlinear Diffusion. Journal Club Presentation. Xiaowei Zhou
1 / 41 Journal Club Presentation Xiaowei Zhou Department of Electronic and Computer Engineering The Hong Kong University of Science and Technology 2009-12-11 2 / 41 Outline 1 Motivation Diffusion process
More informationThis paper is not to be removed from the Examination Halls
~~ST104B ZA d0 This paper is not to be removed from the Examination Halls UNIVERSITY OF LONDON ST104B ZB BSc degrees and Diplomas for Graduates in Economics, Management, Finance and the Social Sciences,
More informationEcon671 Factor Models: Principal Components
Econ671 Factor Models: Principal Components Jun YU April 8, 2016 Jun YU () Econ671 Factor Models: Principal Components April 8, 2016 1 / 59 Factor Models: Principal Components Learning Objectives 1. Show
More informationBig Data Analytics. Lucas Rego Drumond
Big Data Analytics Lucas Rego Drumond Information Systems and Machine Learning Lab (ISMLL) Institute of Computer Science University of Hildesheim, Germany Predictive Models Predictive Models 1 / 34 Outline
More informationPredicting Mutual Fund Performance
Predicting Mutual Fund Performance Oxford, July-August 2013 Allan Timmermann 1 1 UC San Diego, CEPR, CREATES Timmermann (UCSD) Predicting fund performance July 29 - August 2, 2013 1 / 51 1 Basic Performance
More informationProbabilities & Statistics Revision
Probabilities & Statistics Revision Christopher Ting Christopher Ting http://www.mysmu.edu/faculty/christophert/ : christopherting@smu.edu.sg : 6828 0364 : LKCSB 5036 January 6, 2017 Christopher Ting QF
More informationEM-algorithm for Training of State-space Models with Application to Time Series Prediction
EM-algorithm for Training of State-space Models with Application to Time Series Prediction Elia Liitiäinen, Nima Reyhani and Amaury Lendasse Helsinki University of Technology - Neural Networks Research
More informationTrend and long-run relations in electricity prices
Trend and long-run relations in electricity prices Why pre-filtering is inevitable Matteo Pelagatti 2 jointly with A. Gianfreda 1 L. Parisio 2 P. Maranzano 2 1 Free University of Bozen 2 University of
More informationDynamic Pricing for Non-Perishable Products with Demand Learning
Dynamic Pricing for Non-Perishable Products with Demand Learning Victor F. Araman Stern School of Business New York University René A. Caldentey DIMACS Workshop on Yield Management and Dynamic Pricing
More informationEmpirical Market Microstructure Analysis (EMMA)
Empirical Market Microstructure Analysis (EMMA) Lecture 3: Statistical Building Blocks and Econometric Basics Prof. Dr. Michael Stein michael.stein@vwl.uni-freiburg.de Albert-Ludwigs-University of Freiburg
More informationEstimation for the standard and geometric telegraph process. Stefano M. Iacus. (SAPS VI, Le Mans 21-March-2007)
Estimation for the standard and geometric telegraph process Stefano M. Iacus University of Milan(Italy) (SAPS VI, Le Mans 21-March-2007) 1 1. Telegraph process Consider a particle moving on the real line
More informationAuxiliary Particle Methods
Auxiliary Particle Methods Perspectives & Applications Adam M. Johansen 1 adam.johansen@bristol.ac.uk Oxford University Man Institute 29th May 2008 1 Collaborators include: Arnaud Doucet, Nick Whiteley
More informationSimulation and Parametric Estimation of SDEs
Simulation and Parametric Estimation of SDEs Jhonny Gonzalez School of Mathematics The University of Manchester Magical books project August 23, 2012 Motivation The problem Simulation of SDEs SDEs driven
More informationSmall Open Economy RBC Model Uribe, Chapter 4
Small Open Economy RBC Model Uribe, Chapter 4 1 Basic Model 1.1 Uzawa Utility E 0 t=0 θ t U (c t, h t ) θ 0 = 1 θ t+1 = β (c t, h t ) θ t ; β c < 0; β h > 0. Time-varying discount factor With a constant
More informationChapter 6: Derivative-Based. optimization 1
Chapter 6: Derivative-Based Optimization Introduction (6. Descent Methods (6. he Method of Steepest Descent (6.3 Newton s Methods (NM (6.4 Step Size Determination (6.5 Nonlinear Least-Squares Problems
More informationProbabilistic Inference : Changepoints and Cointegration
Probabilistic Inference : Changepoints and Cointegration Chris Bracegirdle c.bracegirdle@cs.ucl.ac.uk 26 th April 2013 Outline 1 Changepoints Modelling Approaches Reset Models : Probabilistic Inferece
More informationEconophysics III: Financial Correlations and Portfolio Optimization
FAKULTÄT FÜR PHYSIK Econophysics III: Financial Correlations and Portfolio Optimization Thomas Guhr Let s Face Chaos through Nonlinear Dynamics, Maribor 21 Outline Portfolio optimization is a key issue
More informationIntroduction to Econometrics Midterm Examination Fall 2005 Answer Key
Introduction to Econometrics Midterm Examination Fall 2005 Answer Key Please answer all of the questions and show your work Clearly indicate your final answer to each question If you think a question is
More informationTopic 12 Overview of Estimation
Topic 12 Overview of Estimation Classical Statistics 1 / 9 Outline Introduction Parameter Estimation Classical Statistics Densities and Likelihoods 2 / 9 Introduction In the simplest possible terms, the
More informationIE598 Big Data Optimization Introduction
IE598 Big Data Optimization Introduction Instructor: Niao He Jan 17, 2018 1 A little about me Assistant Professor, ISE & CSL UIUC, 2016 Ph.D. in Operations Research, M.S. in Computational Sci. & Eng. Georgia
More informationLinear Models in Machine Learning
CS540 Intro to AI Linear Models in Machine Learning Lecturer: Xiaojin Zhu jerryzhu@cs.wisc.edu We briefly go over two linear models frequently used in machine learning: linear regression for, well, regression,
More informationA Modified Fractionally Co-integrated VAR for Predicting Returns
A Modified Fractionally Co-integrated VAR for Predicting Returns Xingzhi Yao Marwan Izzeldin Department of Economics, Lancaster University 13 December 215 Yao & Izzeldin (Lancaster University) CFE (215)
More informationQuick Review on Linear Multiple Regression
Quick Review on Linear Multiple Regression Mei-Yuan Chen Department of Finance National Chung Hsing University March 6, 2007 Introduction for Conditional Mean Modeling Suppose random variables Y, X 1,
More informationSupplement to Quantile-Based Nonparametric Inference for First-Price Auctions
Supplement to Quantile-Based Nonparametric Inference for First-Price Auctions Vadim Marmer University of British Columbia Artyom Shneyerov CIRANO, CIREQ, and Concordia University August 30, 2010 Abstract
More informationWeek 3: Simple Linear Regression
Week 3: Simple Linear Regression Marcelo Coca Perraillon University of Colorado Anschutz Medical Campus Health Services Research Methods I HSMP 7607 2017 c 2017 PERRAILLON ALL RIGHTS RESERVED 1 Outline
More informationEstimating Covariance Using Factorial Hidden Markov Models
Estimating Covariance Using Factorial Hidden Markov Models João Sedoc 1,2 with: Jordan Rodu 3, Lyle Ungar 1, Dean Foster 1 and Jean Gallier 1 1 University of Pennsylvania Philadelphia, PA joao@cis.upenn.edu
More informationTopic 16 Interval Estimation
Topic 16 Interval Estimation Additional Topics 1 / 9 Outline Linear Regression Interpretation of the Confidence Interval 2 / 9 Linear Regression For ordinary linear regression, we have given least squares
More informationMFE Financial Econometrics 2018 Final Exam Model Solutions
MFE Financial Econometrics 2018 Final Exam Model Solutions Tuesday 12 th March, 2019 1. If (X, ε) N (0, I 2 ) what is the distribution of Y = µ + β X + ε? Y N ( µ, β 2 + 1 ) 2. What is the Cramer-Rao lower
More informationPerception: objects in the environment
Zsolt Vizi, Ph.D. 2018 Self-driving cars Sensor fusion: one categorization Type 1: low-level/raw data fusion combining several sources of raw data to produce new data that is expected to be more informative
More informationMulti-armed bandit models: a tutorial
Multi-armed bandit models: a tutorial CERMICS seminar, March 30th, 2016 Multi-Armed Bandit model: general setting K arms: for a {1,..., K}, (X a,t ) t N is a stochastic process. (unknown distributions)
More informationSIMON FRASER UNIVERSITY School of Engineering Science
SIMON FRASER UNIVERSITY School of Engineering Science Course Outline ENSC 810-3 Digital Signal Processing Calendar Description This course covers advanced digital signal processing techniques. The main
More informationParticle Filtering in High-Frequency data.
Particle Filtering in High-Frequency data. A. Platania Facoltá di Scienze Statistiche ed Economiche Via Cesare Battisti, 24 352 Padova Italy platania@stat.unipd.it L.C.G. Rogers Statistical Laboratory
More informationSpeculation and the Bond Market: An Empirical No-arbitrage Framework
Online Appendix to the paper Speculation and the Bond Market: An Empirical No-arbitrage Framework October 5, 2015 Part I: Maturity specific shocks in affine and equilibrium models This Appendix present
More informationOnline Natural Gradient as a Kalman Filter
Online Natural Gradient as a Kalman Filter Yann Ollivier Abstract We cast Amari s natural gradient in statistical learning as a specific case of Kalman filtering. Namely, applying an extended Kalman filter
More informationLecture 5: GPs and Streaming regression
Lecture 5: GPs and Streaming regression Gaussian Processes Information gain Confidence intervals COMP-652 and ECSE-608, Lecture 5 - September 19, 2017 1 Recall: Non-parametric regression Input space X
More informationSociology 6Z03 Review II
Sociology 6Z03 Review II John Fox McMaster University Fall 2016 John Fox (McMaster University) Sociology 6Z03 Review II Fall 2016 1 / 35 Outline: Review II Probability Part I Sampling Distributions Probability
More informationAn Introduction to Malliavin calculus and its applications
An Introduction to Malliavin calculus and its applications Lecture 3: Clark-Ocone formula David Nualart Department of Mathematics Kansas University University of Wyoming Summer School 214 David Nualart
More informationForecasting with ARMA Models
LECTURE 4 Forecasting with ARMA Models Minumum Mean-Square Error Prediction Imagine that y(t) is a stationary stochastic process with E{y(t)} = 0. We may be interested in predicting values of this process
More informationOptimization. Benjamin Recht University of California, Berkeley Stephen Wright University of Wisconsin-Madison
Optimization Benjamin Recht University of California, Berkeley Stephen Wright University of Wisconsin-Madison optimization () cost constraints might be too much to cover in 3 hours optimization (for big
More informationWeek 1 Quantitative Analysis of Financial Markets Distributions A
Week 1 Quantitative Analysis of Financial Markets Distributions A Christopher Ting http://www.mysmu.edu/faculty/christophert/ Christopher Ting : christopherting@smu.edu.sg : 6828 0364 : LKCSB 5036 October
More informationMalliavin 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 informationAn Evolving Gradient Resampling Method for Machine Learning. Jorge Nocedal
An Evolving Gradient Resampling Method for Machine Learning Jorge Nocedal Northwestern University NIPS, Montreal 2015 1 Collaborators Figen Oztoprak Stefan Solntsev Richard Byrd 2 Outline 1. How to improve
More informationApproximate Bayesian inference
Approximate Bayesian inference Variational and Monte Carlo methods Christian A. Naesseth 1 Exchange rate data 0 20 40 60 80 100 120 Month Image data 2 1 Bayesian inference 2 Variational inference 3 Stochastic
More informationFinal Exam November 24, Problem-1: Consider random walk with drift plus a linear time trend: ( t
Problem-1: Consider random walk with drift plus a linear time trend: y t = c + y t 1 + δ t + ϵ t, (1) where {ϵ t } is white noise with E[ϵ 2 t ] = σ 2 >, and y is a non-stochastic initial value. (a) Show
More informationProblems. Suppose both models are fitted to the same data. Show that SS Res, A SS Res, B
Simple Linear Regression 35 Problems 1 Consider a set of data (x i, y i ), i =1, 2,,n, and the following two regression models: y i = β 0 + β 1 x i + ε, (i =1, 2,,n), Model A y i = γ 0 + γ 1 x i + γ 2
More informationModel-Based Diagnosis of Chaotic Vibration Signals
Model-Based Diagnosis of Chaotic Vibration Signals Ihab Wattar ABB Automation 29801 Euclid Ave., MS. 2F8 Wickliffe, OH 44092 and Department of Electrical and Computer Engineering Cleveland State University,
More informationGeneralized Hypothesis Testing and Maximizing the Success Probability in Financial Markets
Generalized Hypothesis Testing and Maximizing the Success Probability in Financial Markets Tim Leung 1, Qingshuo Song 2, and Jie Yang 3 1 Columbia University, New York, USA; leung@ieor.columbia.edu 2 City
More informationTutorial: PART 1. Online Convex Optimization, A Game- Theoretic Approach to Learning.
Tutorial: PART 1 Online Convex Optimization, A Game- Theoretic Approach to Learning http://www.cs.princeton.edu/~ehazan/tutorial/tutorial.htm Elad Hazan Princeton University Satyen Kale Yahoo Research
More informationONLINE LEARNING WITH KERNELS: OVERCOMING THE GROWING SUM PROBLEM. Abhishek Singh, Narendra Ahuja and Pierre Moulin
22 IEEE INTERNATIONAL WORKSHOP ON MACHINE LEARNING FOR SIGNAL PROCESSING, SEPT. 23 26, 22, SANTANDER, SPAIN ONLINE LEARNING WITH KERNELS: OVERCOMING THE GROWING SUM PROBLEM Abhishek Singh, Narendra Ahuja
More informationFinancial Risk and Returns Prediction with Modular Networked Learning
arxiv:1806.05876v1 [cs.lg] 15 Jun 2018 Financial Risk and Returns Prediction with Modular Networked Learning Carlos Pedro Gonçalves June 18, 2018 University of Lisbon, Instituto Superior de Ciências Sociais
More informationMarket-making with Search and Information Frictions
Market-making with Search and Information Frictions Benjamin Lester Philadelphia Fed Venky Venkateswaran NYU Stern Ali Shourideh Carnegie Mellon University Ariel Zetlin-Jones Carnegie Mellon University
More informationInformation Choice in Macroeconomics and Finance.
Information Choice in Macroeconomics and Finance. Laura Veldkamp New York University, Stern School of Business, CEPR and NBER Spring 2009 1 Veldkamp What information consumes is rather obvious: It consumes
More informationUnconstrained minimization of smooth functions
Unconstrained minimization of smooth functions We want to solve min x R N f(x), where f is convex. In this section, we will assume that f is differentiable (so its gradient exists at every point), and
More informationSequential Monte Carlo Methods for Bayesian Computation
Sequential Monte Carlo Methods for Bayesian Computation A. Doucet Kyoto Sept. 2012 A. Doucet (MLSS Sept. 2012) Sept. 2012 1 / 136 Motivating Example 1: Generic Bayesian Model Let X be a vector parameter
More informationThe Real Option Approach to Plant Valuation: from Spread Options to Optimal Switching
The Real Option Approach to Plant Valuation: from Spread Options to René 1 and Michael Ludkovski 2 1 Bendheim Center for Finance Department of Operations Research & Financial Engineering Princeton University
More informationThomas Knispel Leibniz Universität Hannover
Optimal long term investment under model ambiguity Optimal long term investment under model ambiguity homas Knispel Leibniz Universität Hannover knispel@stochastik.uni-hannover.de AnStAp0 Vienna, July
More informationMultivariate GARCH models.
Multivariate GARCH models. Financial market volatility moves together over time across assets and markets. Recognizing this commonality through a multivariate modeling framework leads to obvious gains
More informationQ-Learning and SARSA: Machine learningbased stochastic control approaches for financial trading
Q-Learning and SARSA: Machine learningbased stochastic control approaches for financial trading Marco CORAZZA (corazza@unive.it) Department of Economics Ca' Foscari University of Venice CONFERENCE ON COMPUTATIONAL
More informationMotivation Non-linear Rational Expectations The Permanent Income Hypothesis The Log of Gravity Non-linear IV Estimation Summary.
Econometrics I Department of Economics Universidad Carlos III de Madrid Master in Industrial Economics and Markets Outline Motivation 1 Motivation 2 3 4 5 Motivation Hansen's contributions GMM was developed
More informationForecasting in the presence of recent structural breaks
Forecasting in the presence of recent structural breaks Second International Conference in memory of Carlo Giannini Jana Eklund 1, George Kapetanios 1,2 and Simon Price 1,3 1 Bank of England, 2 Queen Mary
More informationOptimization. Escuela de Ingeniería Informática de Oviedo. (Dpto. de Matemáticas-UniOvi) Numerical Computation Optimization 1 / 30
Optimization Escuela de Ingeniería Informática de Oviedo (Dpto. de Matemáticas-UniOvi) Numerical Computation Optimization 1 / 30 Unconstrained optimization Outline 1 Unconstrained optimization 2 Constrained
More informationTesting for a break in persistence under long-range dependencies and mean shifts
Testing for a break in persistence under long-range dependencies and mean shifts Philipp Sibbertsen and Juliane Willert Institute of Statistics, Faculty of Economics and Management Leibniz Universität
More informationEconomics 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 informationDiscrete Variables and Gradient Estimators
iscrete Variables and Gradient Estimators This assignment is designed to get you comfortable deriving gradient estimators, and optimizing distributions over discrete random variables. For most questions,
More informationA Bayesian Perspective on Residential Demand Response Using Smart Meter Data
A Bayesian Perspective on Residential Demand Response Using Smart Meter Data Datong-Paul Zhou, Maximilian Balandat, and Claire Tomlin University of California, Berkeley [datong.zhou, balandat, tomlin]@eecs.berkeley.edu
More informationAdaptive Filter Theory
0 Adaptive Filter heory Sung Ho Cho Hanyang University Seoul, Korea (Office) +8--0-0390 (Mobile) +8-10-541-5178 dragon@hanyang.ac.kr able of Contents 1 Wiener Filters Gradient Search by Steepest Descent
More informationGaussian with mean ( µ ) and standard deviation ( σ)
Slide from Pieter Abbeel Gaussian with mean ( µ ) and standard deviation ( σ) 10/6/16 CSE-571: Robotics X ~ N( µ, σ ) Y ~ N( aµ + b, a σ ) Y = ax + b + + + + 1 1 1 1 1 1 1 1 1 1, ~ ) ( ) ( ), ( ~ ), (
More informationAsset Pricing. Chapter IX. The Consumption Capital Asset Pricing Model. June 20, 2006
Chapter IX. The Consumption Capital Model June 20, 2006 The Representative Agent Hypothesis and its Notion of Equilibrium 9.2.1 An infinitely lived Representative Agent Avoid terminal period problem Equivalence
More information