Statistical Inference for Ergodic Point Processes and Applications to Limit Order Book

Size: px
Start display at page:

Download "Statistical Inference for Ergodic Point Processes and Applications to Limit Order Book"

Transcription

1 Statistical Inference for Ergodic Point Processes and Applications to Limit Order Book Simon Clinet PhD student under the supervision of Pr. Nakahiro Yoshida Graduate School of Mathematical Sciences, University of Tokyo CREST, Japan Science and Technology Agency Toyama Symposium October 2, Introduction Most financial transactions take place nowadays in electronic markets. Participating to continuous-time double auctions, agents can freely send buying or selling orders at different prices that are automatically matched according precise rules. As this matching process is rather complex and the orders sent by market participants are asynchronous, they are centralized in an Electronic Limit Order Book (also denoted LOB), waiting to be executed according to their price and time priority. A LOB is thus a multidimensional queuing system, each dimension representing a price level, and each queue containing the waiting orders that have not been executed yet, sorted by their arrival time. Agents can then interact with this dynamical system via three elementary mechanisms. They may submit a buying (resp. selling) limit order that will increase the size of one queue on the bid (resp. ask ) side of the LOB. They also may send a buying (resp. selling) market order that will immediately consume the corresponding liquidity at the best available price. Finally they can submit cancellations orders to remove one of their latent limit order in the LOB. Driven by these simple events, the characteristics of the Order Book, such as its mid price, its shape, or the number of orders submitted in a given time window are subject to random fluctuations as time passes. As the macroscopic price movements of an asset are determined by the evolution of its LOB through time, understanding the stochastic structure of this object is a fundamental issue. It is also a way to describe in a high-dimensional context the microstructure phenomena related to the stock price movements. Indeed, phenomena such as the Epps effect, that are traditionally represented by a noise process in many studies that are only interested in the modelling of the price process itself, are captured when the whole limit order book is described. As the availability of high-frequency financial data has made in-depth analysis of LOB dynamics accessible, it is now possible to use statistical tools in order to describe either empirical facts about their shapes and their evolution 14, 4], or to design and select stochastic models that are able to reproduce some of those This work was in part supported by CREST Japan Science and Technology Agency; Japan Society for the Promotion of Science Grants-in-Aid for Scientific Research No (Scientific Research), No (Challenging Exploratory Research); NS Solutions Corporation; and by a Cooperative Research Program of the Institute of Statistical Mathematics. 1

2 stylized facts. For the latter, many studies have been conducted in order to model the stochastic behavior of limit order books, the simplest approaches being zero-intelligence models, in which interarrival times are exponentially distributed. The seminal work about this basic representation 2] has been followed by many extensions, see for example 1, 6, 8, 7, 15]. Lately, more complex dynamics have been studied, including Markovian models 11], Hawkes process driven bid-ask prices 3, 22], or even mixture of those approaches 2]. Since a LOB is mechanically driven by the orders that are submitted through time, many authors choose to see a LOB through the stochastic structure of the interarrival times between two events. In other words, a Limit Order Book is often described as a high-dimensional point process, whose components are integer measures counting the waiting times between two orders of the same type, and at the same price level. In a parametric context, estimating the parameter θ based on the observations is thus a very general and crucial issue that may take place in two distinct asymptotics. As, at least for liquid stocks, a tremendously large number of events happen during short periods of times, the heavy traffic limit (very large number of events on a finite period of time) seems to be a good way to construct consistent estimators. In 17], a sequence of multivariate point processes is thus assumed to be observable on a finite time window. Under suitable assumptions on the sequence of stochastic intensities itself, it is shown that even in this non-ergodic context it is possible to conduct the quasi likelihood analysis procedure (QLA for short). In particular, both quasi maximum likelihood estimator (QMLE) and quasi bayesian estimator (QBE) are consistent and asymptotically mixed normal. On the other hand, in this work, we are interested in the long run characteristics of the LOB seen as a point process. as the time parameter T tends to infinity, assuming that the LOB satisfies suitable ergodicity assumptions, we aim at taking advantage of this regularity in order to derive the asymptotic properties of the QMLE and the QBE. This problematic is of course not new since the consistency and the asymptotic normality of the maximum likelihood estimator for ergodic stationary point processes was shown a few decades ago in 16] and 18]. Furthermore, maximum likelihood estimations have also been empirically conducted for the abovementioned models, but the fact that the point process is ergodic is sometimes unclear. In this article we thus give general ergodicity and regularity assumptions for point processes under which all the results from the general QLA can be derived. In particular, we do not necessarily require the stationarity of the point process. We then give some applications and examples that are picked from the literature, and that verify those general conditions. 2 Multivariate point process Let B = (Ω, F, F, P), F = (F t ) t R+, be a stochastic basis, and assume that we are given a multidimensional point process N t = (N α t ) α I, I = {1,..d}. For the sake of simplicity N is supposed to be defined on R +, N = (N t ) t R+, N =, and its components N α are supposed to have no common jumps. For some finite dimensional compact space Θ R n, we consider the family of increasing processes Λ t (θ) = t λ(s, θ)ds, and we assume that there exists θ Θ such that λ(t, θ ) is the F t -intensity of N t. The process Ñ t = N t Λ t (θ ) is thus a local martingale. All along the paper, if x designates a real number, a vector or a matrix, x = i x i. If X is a random variable, X p = E X p ] 1 p. For a measured space (E, E), C b (E, R) is the set of continuous, bounded functions from (E, E) to (R, B(R)). In order to derive the asymptotic properties of the quasi maximum likelihood estimator (QMLE), we consider the following assumptions : A1] The mapping λ : Ω R + Θ R + is F B(R + ) B(Θ)-measurable. Moreover, almost surely : (i) for any θ Θ, s λ(s, θ) is left continuous. 2

3 (ii) for any s R +, θ λ(s, θ) is in C 2 (Θ). A2] The intensity processes and their first derivatives are non degenerate in the following sense : 2 (i) for any p > 1, sup t R+ i= supθ Θ i θ λ(t, θ) p < + (ii) for any p > 1, sup t R+ sup θ Θ λ(t, θ) 1 1 {λ(t,θ) } p < + (iii) For any θ Θ, λ(t, θ) = iff λ(t, θ ) = A3] Ergodicity. There exists a mapping π : C b (R 4, R) Θ R such that for any (ψ, θ) C b (R 4, R) Θ the following convergence holds : 1 T ψ(λ(s, θ ), λ(s, θ), θ λ(s, θ), T θλ(s, 2 θ))ds P π(ψ, θ) (1) In section 3 we define the likelihood function and the QMLE. We then state its consistency and its asymptotic normality. In section 3.4 we finally derive the convergence of moments of both the QMLE and the QBE under stronger versions of A1]-A4]. 3 Statistical inference 3.1 Definitions and asymptotic properties of the QMLE For any θ Θ, the log-likelihood function can be expressed as (see 9], proposition 7.2.III) : l T (θ) = α I log(λ α (s, θ))dn α s α I λ α (s, θ)ds (2) up to an addition by a term independent of θ. A quasi maximum likelihood estimator ˆθ T asymptotically maximizes the rescaled likelihood function in the following sense: Put, l T (ˆθ T ) T l T (θ) max o P (1) (3) θ Θ T The key lemma to show the consistency of the QMLE is the following Lemma 3.1. There exists Y(θ) such that Y T (θ) = 1 T (l T (θ) l T (θ )) (4) sup Y T (θ) Y(θ) θ Θ The proof of the above lemma makes extensive use of the ergodicity assumption. Indeed, Y has actually the following form : Y(θ) = lim T + 1 T α I ] λ α (s, θ) λ α (s, θ ) log λα (s, θ) λ α (s, θ ) λα (s, θ ) ds (5) To conclude, we need to assume the following non-degeneracy condition on Y : 3

4 A4] For any θ Θ {θ }, Y(θ). Theorem 3.2. Under A4], a QMLE estimator ˆθ T is consistent. ˆθ T P θ The asymptotic normality of the QMLE can be then obtained following the standard procedure. Writing : θ lt L (θ ) = α I and λ α (s, θ ) 1 θ λ α (s, θ )1 {λ α (s,θ ) }dñ α s (6) 2 θl L T (θ) = α I α I + α I It is easy to deduce the existence of the Fisher information : θ ( λ α (s, θ) 1 θ λ α (s, θ) ) 1 {λ α (s,θ ) }dñ α t (7) ( θ λ α ) 2 (s, θ)λ α (s, θ) 2 λ α (s, θ )1 {λ α (s,θ ) }ds (8) 2 θλ α (s, θ)λ α (s, θ) 1 (λ α (s, θ) λ α (s, θ ))1 {λα (s,θ ) }ds (9) Lemma 3.3. There exists Γ R n n such that : Under A1],A2] and A3], if V T is a ball centered on θ shrinking to {θ }, then : sup T 1 θl 2 T (θ) + Γ P θ V T (1) Once again the ergodicity assumption is crucial here since we have the following representation for Γ : 1 Γ = lim T + T α I Finally, thanks to the following lemma Lemma 3.4. we have : ( ) 1 θ l ut (θ ) d Γ 1 2 (Wu ) u,1] ut u,1] ( θ λ α ) 2 (s, θ )λ α (s, θ ) 1 1 {λ α (s,θ ) }ds (11) where W is a standard Brownian motion, and d designs the convergence in distribution of the whole process. This reduces to show a functional central limit theorem. From the previous lemmas we easily deduce the following convergence. Theorem 3.5. Put θ T a QMLE estimator. we have : T (θt θ ) d Γ 1 2 N Where N is a standard normal distribution, and provided that Γ is invertible. 4

5 3.2 QLA and convergence of moments In this section we slightly strengthen the previous assumptions about ergodicity and derivability of the intensity process in order to apply the quasi likelihood analysis and get the convergence of moments of the QMLE as well as the convergence of moments for the QBE. We write C (R 4, R) the set of functions ψ : (x, y, z, t) ψ(x, y, z, t) from R 4 to R of class C 1 such that ψ and its gradient are of polynomial growth in (x, y, z, t, 1 x, 1 y ). A1 ] The mapping λ : Ω R + Θ R + is F B(R + ) B(Θ)-measurable. Moreover, almost surely : (i) for any θ Θ, s λ(s, θ) is left continuous. (ii) for any s R +, θ λ(s, θ) is in C 3 (Θ). A2 ] The intensity processes and their first derivatives are non degenerate in the following sense : 3 (i) for any p > 1, sup t R+ i= supθ Θ θ iλ(t, θ) p < + (ii) for any p > 1, sup t R+ supθ Θ λ(t, θ) 1 1 {λ(t,θ) } p < + (iii) For any θ Θ, λ(t, θ) = iff λ(t, θ ) = A3 ] Ergodicity. There exists a mapping π : C (R 4, R) Θ R and there exists < γ < 1 2 such that for any (ψ, θ) C (R 4, R) Θ the following convergence holds : sup T γ 1 T ψ(λ(s, θ ), λ(s, θ), θ λ(s, θ), 2 θ Θ T θλ(s, θ))ds π(ψ, θ) (12) p A4 ] Define χ as: We assume that χ = inf Y(θ) θ Θ\{θ } θ θ 2 χ > For a given prior density p on the space Θ, we define the quasi Bayesian estimator as follows: ] 1 θ T = exp(l T (θ))p(θ)dθ θexp(l T (θ))p(θ)dθ (13) Θ Θ We can then apply results from 21] about quasi-likelihood analysis based on polynomial type large deviations. To derive our main result about convergence of moments we represent the likelihood process as follows. For any u U T = {u R k θ + T 1/2 u Θ}, we write θ u = θ + T 1/2 u and define the likelihood field : The main result is the following : Z T (u) = exp{l T (θ u ) l T (θ )} Theorem 3.6. Under A1 ]-A4 ], the two following results hold: Polynomial type large deviation inequality for every L >, there exists C L such that : ] P sup Z T (u) e r C L u U T, u >r r L Convergence of moments If ˆθ T is the QMLE and θ T the QBE, we have : E f( ] T (ˆθ T θ )) Ef(Γ 1 2 N )] E f( ] T ( θ T θ )) Ef(Γ 1 2 N )] for any continuous f with polynomial growth. 5

6 4 Applications to Limit Order Book As explained in the introduction, one can always associate to a Limit Order Book process the point process counting that counts the different events that occur through time. If the limit order book is constituted of m N queues stored in the process X t Z m, we can associate m one dimensional counting processes for the limit orders, m for the cancellations, and finally 2 more for the market orders, thus a 2m + 2-dimensional point process (C t, L t, M t ) t. We thus consider a few examples for which the ergodicity condition is satisfied. 4.1 When intensities depend on a Markovian underlying process Some models assume the existence of an underlying process (Y t ) t adapted to the filtration F = (F t ) t R+ that the parametrized intensity processes write as : such λ(t, θ) = h(y t, θ) (14) Example 4.1. In 1], Abergel and Jedidi define a model in which the cancellation intensities are linear functions of the size of the queues of the LOB itself, and other intensities remain constant. In other words, any cancellation intensity at a given level α is of the form : λ C,α (t) = λ C,α X α (t) (15) This feedback enables the authors to show that the vector process containing the sizes of the queues is thus a Markovian V-geometric process (see 13, 12] for a deep insight of this notion). The ergodicity condition A3 ] is thus automatically verified for this simple case. Example 4.2. Another example comes from 19]. In this article, the authors consider that the price process of an asset S t is driven by a Brownian motion. They also assume that the point processes counting the limit orders on the best bid and best ask are Cox processes whose stochastic intensities are functions of the fractional part of S t, Y t := {S t } = S t mod 1. They construct then a non-parametric estimator of the function h using the ergodicity properties of the Markov process Y. It is immediate to see that a parametrized model {h(., θ) θ Θ} would satisfy our ergodicity condition. Example 4.3. Finally, in 11], the model described in 4.1 is generalized. X is defined as a general pure jump Markov process. This, in turn, implies that the intensity vector is a pure function of the LOB state : λ(t, θ) = h(x t, θ) (16) If the intensities of cancellation are sufficiently large when the size of the limits becomes too big, it is shown in 11] that once again the markov Process X is V-geometrically ergodic. 4.2 When intensities follow an exponential Hawkes dynamic In this section we are interested in the special case of multivariate Hawkes process with exponential kernel. Let N t = (Nt α ) α I, I = {1,..d}, N =, be a multidimensional point process and write F N = (Ft N ) t R+, where Ft N = σ{n s s t} is the canonical filtration of N. We say that N is a linear Hawkes process or Hawkes self-exciting process starting from if there exist h : R + R d d + and ν (R +) d such that the Ft N -intensity λ(t) of N writes : 6

7 for any α I. λ α (t) = ν α + β I t h αβ (t s)dn β s Given A = a αβ ] αβ and C = c αβ ] αβ R d d +, we say that N is an exponential Hawkes process if the kernel functions h αβ are of the form : h αβ (s) = c αβ e a αβs Those processes were introduced by Hawkes in 1971, see 1], and were extensively used to model earthquakes and their aftershocks. Lately they have been used in finance to represent the fact that market and limit orders seem to trigger each other, see for example 2]. Let us recall in the following a few results about this family of processes. We define the elementary excitations as : ɛ αβ (t) = t h αβ (t s)dn β s and the matrix E(t) = ] ɛ αβ(t). Finally we define the excitation ratio matrix as Φ = cαβ αβ a αβ ]αβ, and ρ(φ) its spectral radius. We have the following results : Proposition 4.4. Assume that ρ(φ) < 1. Then : (i) There exists a unique stationary point process N defined on the whole real line and on the same probability space as N, such that its F N t -intensity λ verifies : λ α (t) = ν α + β I t h αβ (t s)d N β s (ii) Let S be the shift operator, meaning that for any t R +, S t N = (N s+t ) s R+. Then N is stable in the following sense : S t N D N R+ where D designates the weak convergence associated to the vague topology on integer valued measures. See 5] for deeper explanations about this mode of convergence. Proposition 4.5. Assume that ρ(φ) < 1. Then : (i) The elementary excitation process E is a Markov Process taking values in R d d +. (ii) E is V -geometrically ergodic, and moreover V : R+ d d R + can be chosen as V (ɛ) = e M,ɛ for some M = m αβ ] αβ R d d + and with M, ɛ = α,β I m αβɛ αβ. We now consider an exponential Hawkes process as a model parametrized by the triplet (ν, c, a) R + R d d + R d d +. More precisely we consider a compact parameter state Θ R + R d d + R+ d d such that for any triplet θ = (ν, c, a), for any α, β I, < ν ν α ν < + c c αβ c < + < a a αβ ā < + The following result is a consequence of the previous propositions : 7

8 Theorem 4.6. Assume that ρ(φ) < 1. Then the non-stationary exponential Hawkes process verifies A1 ]-A4 ] and the QLA applies. The main difficulty consists in showing the ergodicity assumption A3 ]. Indeed, given the form of the intensities, the process (λ(s, θ ), λ(s, θ), θ λ(s, θ), θ 2 λ(s, θ)) cannot be written as a pure function of the Markov Process E(t). On the other hand, since the exponential Hawkes intensities are stable and are generated by a fast decreasing kernel, they should be mixing in the following sense : For a given process (X t ) t R+ taking values in some state space E, and a set of functions C from E to R, we say that X is C-mixing if for any φ, ψ C, the quantity ρ u = sup s R + Covφ(X s ), ψ(x s+u )] It turns out that the exponential Hawkes process satisfies the following property, which in turns implies the ergodicity condition A3 ]. B1 ] Mixing (λ(t, θ ), λ(t, θ), θ λ(t, θ), 2 θ λ(t, θ)) t R + is C (R 4, R)-mixing. Moreover the rate ρ verifies : ρ u = o(u ɛ ) for some ɛ > Stability There exists λ such that for any (ψ, θ) C (R 4, R) Θ, 2γ 1 2γ t γ Eψ(λ(t, θ ), λ(t, θ), θ λ(t, θ), 2 θλ(t, θ))] Eψ( λ(θ ), λ(θ), θ λ(θ), 2 θ λ(θ))] In particular, the mapping π writes: π(ψ, θ) = Eψ( λ(θ ), λ(θ), θ λ(θ), 2 θ λ(θ))] A typical example including a Hawkes process is the following : Example 4.7. In 2], the point process associated to the limit and market orders (L t, M t ) t form a stable multivariate exponential Hawkes process. As for 4.1, the cancellation intensities are linear functions of the size of the queues. The ergodicity condition A3 ] is an easy consequence of 4.6. References 1] Abergel, F., Jedidi, A.: A mathematical approach to order book modeling. International Journal of Theoretical and Applied Finance 16(5) (213) 2] Abergel, F., Jedidi, A.: Long time behaviour of a hawkes process-based limit order book. Preprint (215) 3] Bacry, E., Delattre, S., Hoffmann, M.: Modelling microstructure noise with mutually exciting point processes. Quantitative Finance 13(1), (213) 4] Bouchaud, J., Farmer, J., Lillo, F.: How markets slowly digest changes in supply and demand. ArXiv e-prints (28) 5] Bremaud, P., Massoulie, L.: Stability of non-linear hawkes processes. The Annals of Probability 24(3), (1996) 6] Cont, R., de Larrard, A.: Order book dynamics in liquid markets : limit theorems and diffusion approximations. ArXiv e-prints (212) 7] Cont, R., de Larrard, A.: Price dynamics in a markovian limit order market. SIAM Journal on Financial Mathematics (213) 8

9 8] Cont, R., Stoikov, S., Talreja, R.: A stochastic model for order book dynamics. Operations Research 58(3), (21) 9] Daley, D., Vere-Jones, D.: An Introduction to the Theory of Point Processes, vol. 1, 2 edn. Springer (23) 1] Hawkes, A.G.: Point spectra of some mutually exciting point processes. J. R. Stat. Soc. Ser. B Stat. Methodol. 33, (1971) 11] Huang, W., Lehalle, C., Rosenbaum, M.: Simulating and analyzing order book data: The queue-reactive model. SIAM Journal on Financial Mathematics 4(1) (214) 12] Meyn, S., Tweedie, R.: Stability of markovian processes iii. Advances in applied probability 25(3), (1993) 13] Meyn, S., Tweedie, R.: Markov chains and stochastic stability, 2 edn. Cambridge University Press (29) 14] Mike, S., Farmer, J.: An empirical behavioral model of liquidity and volatility. Journal of Economics Dynamics and Control 32, (28) 15] Muni Toke, I.: The order book as a queing system : average depth and influence of the size of limit orders. arxiv: (213) 16] Ogata, Y.: The asymptotic behaviour of maximum likelihood estimators for stationary point processes. Annals of the Institute of Statistical Mathematics 3(A), (1978) 17] Ogihara, T., Yoshida, N.: Quasi likelihood analysis for point process regression models. preprint (215) 18] Puri, M.L., Tuan, P.D.: Maximum likelihood estimation for stationary point process. Proceedings of the National Academy of Sciences 83, (1986) 19] Rosenbaum, M., Delattre, S., Robert, C.Y.: Estimating the efficient price from the order flow : a brownian cox process approach. arxiv: (1971) 2] Smith, E., Farmer, J., Gillemot, L., Krishnamurthy, S.: Statistical theory of the continuous double auction. Quantitative Finance 3, (23) 21] Yoshida, N.: Polynomial type large deviation inequalities and quasi-likelihood inequalities for stochastic differential equations. Springer (27) 22] Zheng, B., Roueff, F., Abergel, F.: Modelling bid and ask prices using constrained hawkes processes. ergodicity and saling limit. SIAM Journal on Financial Mathematics 5(1), (214) 9

Statistical inference for ergodic point processes and application to Limit Order Book

Statistical inference for ergodic point processes and application to Limit Order Book Available online at www.sciencedirect.com ScienceDirect Stochastic Processes and their Applications 127 (217) 18 1839 www.elsevier.com/locate/spa Statistical inference for ergodic point processes and application

More information

Modelling Trades-Through in a Limit Order Book Using Hawkes Processes

Modelling Trades-Through in a Limit Order Book Using Hawkes Processes Vol. 6, 212-22 June 14, 212 http://dx.doi.org/1.518/economics-ejournal.ja.212-22 Modelling Trades-Through in a Limit Order Book Using Hawkes Processes Ioane Muni Toke Ecole Centrale Paris and University

More information

Order book modeling and market making under uncertainty.

Order book modeling and market making under uncertainty. Order book modeling and market making under uncertainty. Sidi Mohamed ALY Lund University sidi@maths.lth.se (Joint work with K. Nyström and C. Zhang, Uppsala University) Le Mans, June 29, 2016 1 / 22 Outline

More information

Estimating the efficient price from the order flow : a Brownian Cox process approach

Estimating the efficient price from the order flow : a Brownian Cox process approach Estimating the efficient price from the order flow : a Brownian Cox process approach - Sylvain DELARE Université Paris Diderot, LPMA) - Christian ROBER Université Lyon 1, Laboratoire SAF) - Mathieu ROSENBAUM

More information

Some Applications of Hawkes Processes for Order Book Modelling

Some Applications of Hawkes Processes for Order Book Modelling Some Applications of Hawkes Processes for Order Book Modelling Ioane Muni Toke Ecole Centrale Paris Chair of Quantitative Finance First Workshop on Quantitative Finance and Economics International Christian

More information

Analyzing order flows in limit order books with ratios of Cox-type intensities

Analyzing order flows in limit order books with ratios of Cox-type intensities Analyzing order flows in limit order books with ratios of Cox-type intensities Ioane Muni Toke 1 and Nakahiro Yoshida 2 1 Mathématiques et Informatique pour la Complexité et les Systèmes, CentraleSupélec,

More information

Dynamics of Order Positions and Related Queues in a LOB. Xin Guo UC Berkeley

Dynamics of Order Positions and Related Queues in a LOB. Xin Guo UC Berkeley Dynamics of Order Positions and Related Queues in a LOB Xin Guo UC Berkeley Angers, France 09/01/15 Joint work with R. Zhao (UC Berkeley) and L. J. Zhu (U. Minnesota) Outline 1 Background/Motivation 2

More information

Point Process Control

Point Process Control Point Process Control The following note is based on Chapters I, II and VII in Brémaud s book Point Processes and Queues (1981). 1 Basic Definitions Consider some probability space (Ω, F, P). A real-valued

More information

Liquidation in Limit Order Books. LOBs with Controlled Intensity

Liquidation in Limit Order Books. LOBs with Controlled Intensity Limit Order Book Model Power-Law Intensity Law Exponential Decay Order Books Extensions Liquidation in Limit Order Books with Controlled Intensity Erhan and Mike Ludkovski University of Michigan and UCSB

More information

Stability and price scaling limit of a Hawkes-process based order book model

Stability and price scaling limit of a Hawkes-process based order book model Stability and price scaling limit of a Hawkes-process based order book model Aymen Jedidi, Frédéric Abergel To cite this version: Aymen Jedidi, Frédéric Abergel. Stability and price scaling limit of a

More information

Multivariate Hawkes Processes and Their Simulations

Multivariate Hawkes Processes and Their Simulations Multivariate Hawkes Processes and Their Simulations Yuanda Chen September, 2016 Abstract In this article we will extend our discussion to the multivariate Hawkes processes, which are mutually exciting

More information

Least Squares Estimators for Stochastic Differential Equations Driven by Small Lévy Noises

Least Squares Estimators for Stochastic Differential Equations Driven by Small Lévy Noises Least Squares Estimators for Stochastic Differential Equations Driven by Small Lévy Noises Hongwei Long* Department of Mathematical Sciences, Florida Atlantic University, Boca Raton Florida 33431-991,

More information

arxiv: v2 [q-fin.tr] 11 Apr 2013

arxiv: v2 [q-fin.tr] 11 Apr 2013 Estimating the efficient price from the order flow: a Brownian Cox process approach arxiv:131.3114v2 [q-fin.r] 11 Apr 213 Sylvain Delattre LPMA, Université Paris Diderot Paris 7) delattre@math.univ-paris-diderot.fr

More information

Likelihood Function for Multivariate Hawkes Processes

Likelihood Function for Multivariate Hawkes Processes Lielihood Function for Multivariate Hawes Processes Yuanda Chen January, 6 Abstract In this article we discuss the lielihood function for an M-variate Hawes process and derive the explicit formula for

More information

Limit theorems for Hawkes processes

Limit theorems for Hawkes processes Limit theorems for nearly unstable and Mathieu Rosenbaum Séminaire des doctorants du CMAP Vendredi 18 Octobre 213 Limit theorems for 1 Denition and basic properties Cluster representation of Correlation

More information

ECONOMICS 7200 MODERN TIME SERIES ANALYSIS Econometric Theory and Applications

ECONOMICS 7200 MODERN TIME SERIES ANALYSIS Econometric Theory and Applications ECONOMICS 7200 MODERN TIME SERIES ANALYSIS Econometric Theory and Applications Yongmiao Hong Department of Economics & Department of Statistical Sciences Cornell University Spring 2019 Time and uncertainty

More information

Understanding Regressions with Observations Collected at High Frequency over Long Span

Understanding Regressions with Observations Collected at High Frequency over Long Span Understanding Regressions with Observations Collected at High Frequency over Long Span Yoosoon Chang Department of Economics, Indiana University Joon Y. Park Department of Economics, Indiana University

More information

Filtrations, Markov Processes and Martingales. Lectures on Lévy Processes and Stochastic Calculus, Braunschweig, Lecture 3: The Lévy-Itô Decomposition

Filtrations, Markov Processes and Martingales. Lectures on Lévy Processes and Stochastic Calculus, Braunschweig, Lecture 3: The Lévy-Itô Decomposition Filtrations, Markov Processes and Martingales Lectures on Lévy Processes and Stochastic Calculus, Braunschweig, Lecture 3: The Lévy-Itô Decomposition David pplebaum Probability and Statistics Department,

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

arxiv: v1 [q-fin.rm] 27 Jun 2017

arxiv: v1 [q-fin.rm] 27 Jun 2017 Risk Model Based on General Compound Hawkes Process Anatoliy Swishchuk 1 2 arxiv:1706.09038v1 [q-fin.rm] 27 Jun 2017 Abstract: In this paper, we introduce a new model for the risk process based on general

More information

Tyler Hofmeister. University of Calgary Mathematical and Computational Finance Laboratory

Tyler Hofmeister. University of Calgary Mathematical and Computational Finance Laboratory JUMP PROCESSES GENERALIZING STOCHASTIC INTEGRALS WITH JUMPS Tyler Hofmeister University of Calgary Mathematical and Computational Finance Laboratory Overview 1. General Method 2. Poisson Processes 3. Diffusion

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

Local consistency of Markov chain Monte Carlo methods

Local consistency of Markov chain Monte Carlo methods Ann Inst Stat Math (2014) 66:63 74 DOI 10.1007/s10463-013-0403-3 Local consistency of Markov chain Monte Carlo methods Kengo Kamatani Received: 12 January 2012 / Revised: 8 March 2013 / Published online:

More information

Operations Research Letters. Instability of FIFO in a simple queueing system with arbitrarily low loads

Operations Research Letters. Instability of FIFO in a simple queueing system with arbitrarily low loads Operations Research Letters 37 (2009) 312 316 Contents lists available at ScienceDirect Operations Research Letters journal homepage: www.elsevier.com/locate/orl Instability of FIFO in a simple queueing

More information

Location Multiplicative Error Model. Asymptotic Inference and Empirical Analysis

Location Multiplicative Error Model. Asymptotic Inference and Empirical Analysis : Asymptotic Inference and Empirical Analysis Qian Li Department of Mathematics and Statistics University of Missouri-Kansas City ql35d@mail.umkc.edu October 29, 2015 Outline of Topics Introduction GARCH

More information

Locally stationary Hawkes processes

Locally stationary Hawkes processes Locally stationary Hawkes processes François Roue LTCI, CNRS, Télécom ParisTech, Université Paris-Saclay Talk based on Roue, von Sachs, and Sansonnet [2016] 1 / 40 Outline Introduction Non-stationary Hawkes

More information

The Asymptotic Theory of Transaction Costs

The Asymptotic Theory of Transaction Costs The Asymptotic Theory of Transaction Costs Lecture Notes by Walter Schachermayer Nachdiplom-Vorlesung, ETH Zürich, WS 15/16 1 Models on Finite Probability Spaces In this section we consider a stock price

More information

ABC methods for phase-type distributions with applications in insurance risk problems

ABC methods for phase-type distributions with applications in insurance risk problems ABC methods for phase-type with applications problems Concepcion Ausin, Department of Statistics, Universidad Carlos III de Madrid Joint work with: Pedro Galeano, Universidad Carlos III de Madrid Simon

More information

Hawkes Processes and their Applications in Finance and Insurance

Hawkes Processes and their Applications in Finance and Insurance Hawkes Processes and their Applications in Finance and Insurance Anatoliy Swishchuk University of Calgary Calgary, Alberta, Canada Hawks Seminar Talk Dept. of Math. & Stat. Calgary, Canada May 9th, 2018

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

Electronic Market Making and Latency

Electronic Market Making and Latency Electronic Market Making and Latency Xuefeng Gao 1, Yunhan Wang 2 June 15, 2018 Abstract This paper studies the profitability of market making strategies and the impact of latency on electronic market

More information

UNCERTAINTY FUNCTIONAL DIFFERENTIAL EQUATIONS FOR FINANCE

UNCERTAINTY FUNCTIONAL DIFFERENTIAL EQUATIONS FOR FINANCE Surveys in Mathematics and its Applications ISSN 1842-6298 (electronic), 1843-7265 (print) Volume 5 (2010), 275 284 UNCERTAINTY FUNCTIONAL DIFFERENTIAL EQUATIONS FOR FINANCE Iuliana Carmen Bărbăcioru Abstract.

More information

arxiv: v1 [q-fin.tr] 17 Aug 2016

arxiv: v1 [q-fin.tr] 17 Aug 2016 General Semi-Markov Model for Limit Order Books: Theory, Implementation and Numerics arxiv:168.56v1 [q-fin.tr] 17 Aug 216 Anatoliy Swishchuk, Katharina Cera, Julia Schmidt and Tyler Hofmeister August 216

More information

Introduction General Framework Toy models Discrete Markov model Data Analysis Conclusion. The Micro-Price. Sasha Stoikov. Cornell University

Introduction 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 information

Random Times and Their Properties

Random Times and Their Properties Chapter 6 Random Times and Their Properties Section 6.1 recalls the definition of a filtration (a growing collection of σ-fields) and of stopping times (basically, measurable random times). Section 6.2

More information

Positive Harris Recurrence and Diffusion Scale Analysis of a Push Pull Queueing Network. Haifa Statistics Seminar May 5, 2008

Positive Harris Recurrence and Diffusion Scale Analysis of a Push Pull Queueing Network. Haifa Statistics Seminar May 5, 2008 Positive Harris Recurrence and Diffusion Scale Analysis of a Push Pull Queueing Network Yoni Nazarathy Gideon Weiss Haifa Statistics Seminar May 5, 2008 1 Outline 1 Preview of Results 2 Introduction Queueing

More information

Some functional (Hölderian) limit theorems and their applications (II)

Some functional (Hölderian) limit theorems and their applications (II) Some functional (Hölderian) limit theorems and their applications (II) Alfredas Račkauskas Vilnius University Outils Statistiques et Probabilistes pour la Finance Université de Rouen June 1 5, Rouen (Rouen

More information

A one-level limit order book model with memory and variable spread

A one-level limit order book model with memory and variable spread A one-level limit order book model with memory and variable spread Jonathan A. Chávez-Casillas José E. Figueroa-López September 8, 216 Abstract Motivated by Cont and Larrard (213) s seminal Limit Order

More information

Tasmanian School of Business & Economics Economics & Finance Seminar Series 1 February 2016

Tasmanian School of Business & Economics Economics & Finance Seminar Series 1 February 2016 P A R X (PARX), US A A G C D K A R Gruppo Bancario Credito Valtellinese, University of Bologna, University College London and University of Copenhagen Tasmanian School of Business & Economics Economics

More information

Functional Limit theorems for the quadratic variation of a continuous time random walk and for certain stochastic integrals

Functional Limit theorems for the quadratic variation of a continuous time random walk and for certain stochastic integrals Functional Limit theorems for the quadratic variation of a continuous time random walk and for certain stochastic integrals Noèlia Viles Cuadros BCAM- Basque Center of Applied Mathematics with Prof. Enrico

More information

A Mathematical Approach to Order Book Modeling

A Mathematical Approach to Order Book Modeling A Mathematical Approach to Order Book Modeling Frédéric Abergel and Aymen Jedidi November, 212 Abstract Motivated by the desire to bridge the gap between the microscopic description of price formation

More information

Order book resilience, price manipulation, and the positive portfolio problem

Order book resilience, price manipulation, and the positive portfolio problem Order book resilience, price manipulation, and the positive portfolio problem Alexander Schied Mannheim University Workshop on New Directions in Financial Mathematics Institute for Pure and Applied Mathematics,

More information

STA205 Probability: Week 8 R. Wolpert

STA205 Probability: Week 8 R. Wolpert INFINITE COIN-TOSS AND THE LAWS OF LARGE NUMBERS The traditional interpretation of the probability of an event E is its asymptotic frequency: the limit as n of the fraction of n repeated, similar, and

More information

On the Goodness-of-Fit Tests for Some Continuous Time Processes

On the Goodness-of-Fit Tests for Some Continuous Time Processes On the Goodness-of-Fit Tests for Some Continuous Time Processes Sergueï Dachian and Yury A. Kutoyants Laboratoire de Mathématiques, Université Blaise Pascal Laboratoire de Statistique et Processus, Université

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

ON THE REGULARITY OF SAMPLE PATHS OF SUB-ELLIPTIC DIFFUSIONS ON MANIFOLDS

ON THE REGULARITY OF SAMPLE PATHS OF SUB-ELLIPTIC DIFFUSIONS ON MANIFOLDS Bendikov, A. and Saloff-Coste, L. Osaka J. Math. 4 (5), 677 7 ON THE REGULARITY OF SAMPLE PATHS OF SUB-ELLIPTIC DIFFUSIONS ON MANIFOLDS ALEXANDER BENDIKOV and LAURENT SALOFF-COSTE (Received March 4, 4)

More information

E-Companion to Fully Sequential Procedures for Large-Scale Ranking-and-Selection Problems in Parallel Computing Environments

E-Companion to Fully Sequential Procedures for Large-Scale Ranking-and-Selection Problems in Parallel Computing Environments E-Companion to Fully Sequential Procedures for Large-Scale Ranking-and-Selection Problems in Parallel Computing Environments Jun Luo Antai College of Economics and Management Shanghai Jiao Tong University

More information

Consistency of the maximum likelihood estimator for general hidden Markov models

Consistency of the maximum likelihood estimator for general hidden Markov models Consistency of the maximum likelihood estimator for general hidden Markov models Jimmy Olsson Centre for Mathematical Sciences Lund University Nordstat 2012 Umeå, Sweden Collaborators Hidden Markov models

More information

Existence, Uniqueness and Stability of Invariant Distributions in Continuous-Time Stochastic Models

Existence, Uniqueness and Stability of Invariant Distributions in Continuous-Time Stochastic Models Existence, Uniqueness and Stability of Invariant Distributions in Continuous-Time Stochastic Models Christian Bayer and Klaus Wälde Weierstrass Institute for Applied Analysis and Stochastics and University

More information

Jump-type Levy Processes

Jump-type Levy Processes Jump-type Levy Processes Ernst Eberlein Handbook of Financial Time Series Outline Table of contents Probabilistic Structure of Levy Processes Levy process Levy-Ito decomposition Jump part Probabilistic

More information

Short-time expansions for close-to-the-money options under a Lévy jump model with stochastic volatility

Short-time expansions for close-to-the-money options under a Lévy jump model with stochastic volatility Short-time expansions for close-to-the-money options under a Lévy jump model with stochastic volatility José Enrique Figueroa-López 1 1 Department of Statistics Purdue University Statistics, Jump Processes,

More information

Practical conditions on Markov chains for weak convergence of tail empirical processes

Practical conditions on Markov chains for weak convergence of tail empirical processes Practical conditions on Markov chains for weak convergence of tail empirical processes Olivier Wintenberger University of Copenhagen and Paris VI Joint work with Rafa l Kulik and Philippe Soulier Toronto,

More information

Exponential martingales: uniform integrability results and applications to point processes

Exponential martingales: uniform integrability results and applications to point processes Exponential martingales: uniform integrability results and applications to point processes Alexander Sokol Department of Mathematical Sciences, University of Copenhagen 26 September, 2012 1 / 39 Agenda

More information

Temporal point processes: the conditional intensity function

Temporal point processes: the conditional intensity function Temporal point processes: the conditional intensity function Jakob Gulddahl Rasmussen December 21, 2009 Contents 1 Introduction 2 2 Evolutionary point processes 2 2.1 Evolutionarity..............................

More information

Asymptotics for posterior hazards

Asymptotics for posterior hazards Asymptotics for posterior hazards Pierpaolo De Blasi University of Turin 10th August 2007, BNR Workshop, Isaac Newton Intitute, Cambridge, UK Joint work with Giovanni Peccati (Université Paris VI) and

More information

One-level limit order books with sparsity and memory

One-level limit order books with sparsity and memory One-level limit order books with sparsity and memory Jonathan A. Chávez-Casillas José E. Figueroa-López July 1, 14 Abstract Motivated by Cont and Larrard (13) s seminal Limit Order Book (LOB) model, we

More information

SIMILAR MARKOV CHAINS

SIMILAR MARKOV CHAINS SIMILAR MARKOV CHAINS by Phil Pollett The University of Queensland MAIN REFERENCES Convergence of Markov transition probabilities and their spectral properties 1. Vere-Jones, D. Geometric ergodicity in

More information

Asymptotic inference for a nonstationary double ar(1) model

Asymptotic inference for a nonstationary double ar(1) model Asymptotic inference for a nonstationary double ar() model By SHIQING LING and DONG LI Department of Mathematics, Hong Kong University of Science and Technology, Hong Kong maling@ust.hk malidong@ust.hk

More information

On Realized Volatility, Covariance and Hedging Coefficient of the Nikkei-225 Futures with Micro-Market Noise

On Realized Volatility, Covariance and Hedging Coefficient of the Nikkei-225 Futures with Micro-Market Noise On Realized Volatility, Covariance and Hedging Coefficient of the Nikkei-225 Futures with Micro-Market Noise Naoto Kunitomo and Seisho Sato October 22, 2008 Abstract For the estimation problem of the realized

More information

Research Reports on Mathematical and Computing Sciences

Research Reports on Mathematical and Computing Sciences ISSN 1342-2804 Research Reports on Mathematical and Computing Sciences Long-tailed degree distribution of a random geometric graph constructed by the Boolean model with spherical grains Naoto Miyoshi,

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

arxiv: v1 [q-fin.tr] 1 May 2013

arxiv: v1 [q-fin.tr] 1 May 2013 Semi Markov model for market microstructure arxiv:135.15v1 [q-fin.tr 1 May 13 Pietro FODRA Laboratoire de Probabilités et Modèles Aléatoires CNRS, UMR 7599 Université Paris 7 Diderot and EXQIM pietro.fodra91@gmail.com

More information

Central-limit approach to risk-aware Markov decision processes

Central-limit approach to risk-aware Markov decision processes Central-limit approach to risk-aware Markov decision processes Jia Yuan Yu Concordia University November 27, 2015 Joint work with Pengqian Yu and Huan Xu. Inventory Management 1 1 Look at current inventory

More information

Introduction Optimality and Asset Pricing

Introduction Optimality and Asset Pricing Introduction Optimality and Asset Pricing Andrea Buraschi Imperial College Business School October 2010 The Euler Equation Take an economy where price is given with respect to the numéraire, which is our

More information

The Skorokhod reflection problem for functions with discontinuities (contractive case)

The Skorokhod reflection problem for functions with discontinuities (contractive case) The Skorokhod reflection problem for functions with discontinuities (contractive case) TAKIS KONSTANTOPOULOS Univ. of Texas at Austin Revised March 1999 Abstract Basic properties of the Skorokhod reflection

More information

High-dimensional Markov Chain Models for Categorical Data Sequences with Applications Wai-Ki CHING AMACL, Department of Mathematics HKU 19 March 2013

High-dimensional Markov Chain Models for Categorical Data Sequences with Applications Wai-Ki CHING AMACL, Department of Mathematics HKU 19 March 2013 High-dimensional Markov Chain Models for Categorical Data Sequences with Applications Wai-Ki CHING AMACL, Department of Mathematics HKU 19 March 2013 Abstract: Markov chains are popular models for a modelling

More information

ON CONVERGENCE RATES OF GIBBS SAMPLERS FOR UNIFORM DISTRIBUTIONS

ON CONVERGENCE RATES OF GIBBS SAMPLERS FOR UNIFORM DISTRIBUTIONS The Annals of Applied Probability 1998, Vol. 8, No. 4, 1291 1302 ON CONVERGENCE RATES OF GIBBS SAMPLERS FOR UNIFORM DISTRIBUTIONS By Gareth O. Roberts 1 and Jeffrey S. Rosenthal 2 University of Cambridge

More information

Outline. A Central Limit Theorem for Truncating Stochastic Algorithms

Outline. A Central Limit Theorem for Truncating Stochastic Algorithms Outline A Central Limit Theorem for Truncating Stochastic Algorithms Jérôme Lelong http://cermics.enpc.fr/ lelong Tuesday September 5, 6 1 3 4 Jérôme Lelong (CERMICS) Tuesday September 5, 6 1 / 3 Jérôme

More information

Change detection problems in branching processes

Change detection problems in branching processes Change detection problems in branching processes Outline of Ph.D. thesis by Tamás T. Szabó Thesis advisor: Professor Gyula Pap Doctoral School of Mathematics and Computer Science Bolyai Institute, University

More information

Modeling Ultra-High-Frequency Multivariate Financial Data by Monte Carlo Simulation Methods

Modeling Ultra-High-Frequency Multivariate Financial Data by Monte Carlo Simulation Methods Outline Modeling Ultra-High-Frequency Multivariate Financial Data by Monte Carlo Simulation Methods Ph.D. Student: Supervisor: Marco Minozzo Dipartimento di Scienze Economiche Università degli Studi di

More information

Unified Discrete-Time and Continuous-Time Models. and High-Frequency Financial Data

Unified Discrete-Time and Continuous-Time Models. and High-Frequency Financial Data Unified Discrete-Time and Continuous-Time Models and Statistical Inferences for Merged Low-Frequency and High-Frequency Financial Data Donggyu Kim and Yazhen Wang University of Wisconsin-Madison December

More information

Long-Run Covariability

Long-Run Covariability Long-Run Covariability Ulrich K. Müller and Mark W. Watson Princeton University October 2016 Motivation Study the long-run covariability/relationship between economic variables great ratios, long-run Phillips

More information

Statistical inference on Lévy processes

Statistical inference on Lévy processes Alberto Coca Cabrero University of Cambridge - CCA Supervisors: Dr. Richard Nickl and Professor L.C.G.Rogers Funded by Fundación Mutua Madrileña and EPSRC MASDOC/CCA student workshop 2013 26th March Outline

More information

Optimal exit strategies for investment projects. 7th AMaMeF and Swissquote Conference

Optimal exit strategies for investment projects. 7th AMaMeF and Swissquote Conference Optimal exit strategies for investment projects Simone Scotti Université Paris Diderot Laboratoire de Probabilité et Modèles Aléatories Joint work with : Etienne Chevalier, Université d Evry Vathana Ly

More information

An Uncertain Control Model with Application to. Production-Inventory System

An Uncertain Control Model with Application to. Production-Inventory System An Uncertain Control Model with Application to Production-Inventory System Kai Yao 1, Zhongfeng Qin 2 1 Department of Mathematical Sciences, Tsinghua University, Beijing 100084, China 2 School of Economics

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

Potentials of Unbalanced Complex Kinetics Observed in Market Time Series

Potentials of Unbalanced Complex Kinetics Observed in Market Time Series Potentials of Unbalanced Complex Kinetics Observed in Market Time Series Misako Takayasu 1, Takayuki Mizuno 1 and Hideki Takayasu 2 1 Department of Computational Intelligence & Systems Science, Interdisciplinary

More information

Simulation and Calibration of a Fully Bayesian Marked Multidimensional Hawkes Process with Dissimilar Decays

Simulation and Calibration of a Fully Bayesian Marked Multidimensional Hawkes Process with Dissimilar Decays Simulation and Calibration of a Fully Bayesian Marked Multidimensional Hawkes Process with Dissimilar Decays Kar Wai Lim, Young Lee, Leif Hanlen, Hongbiao Zhao Australian National University Data61/CSIRO

More information

Self-exciting point processes with applications in finance and medicine

Self-exciting point processes with applications in finance and medicine MTNS8 28/5/21 Self-exciting point processes with applications in finance and medicine L. Gerencsér, C. Matias, Zs. Vágó, B. Torma and B. Weiss 1 Introduction Stochastic systems driven by point processes

More information

is a Borel subset of S Θ for each c R (Bertsekas and Shreve, 1978, Proposition 7.36) This always holds in practical applications.

is a Borel subset of S Θ for each c R (Bertsekas and Shreve, 1978, Proposition 7.36) This always holds in practical applications. Stat 811 Lecture Notes The Wald Consistency Theorem Charles J. Geyer April 9, 01 1 Analyticity Assumptions Let { f θ : θ Θ } be a family of subprobability densities 1 with respect to a measure µ on a measurable

More information

Covariance function estimation in Gaussian process regression

Covariance function estimation in Gaussian process regression Covariance function estimation in Gaussian process regression François Bachoc Department of Statistics and Operations Research, University of Vienna WU Research Seminar - May 2015 François Bachoc Gaussian

More information

A uniform central limit theorem for neural network based autoregressive processes with applications to change-point analysis

A uniform central limit theorem for neural network based autoregressive processes with applications to change-point analysis A uniform central limit theorem for neural network based autoregressive processes with applications to change-point analysis Claudia Kirch Joseph Tadjuidje Kamgaing March 6, 20 Abstract We consider an

More information

Can we do statistical inference in a non-asymptotic way? 1

Can we do statistical inference in a non-asymptotic way? 1 Can we do statistical inference in a non-asymptotic way? 1 Guang Cheng 2 Statistics@Purdue www.science.purdue.edu/bigdata/ ONR Review Meeting@Duke Oct 11, 2017 1 Acknowledge NSF, ONR and Simons Foundation.

More information

Backward martingale representation and endogenous completeness in finance

Backward martingale representation and endogenous completeness in finance Backward martingale representation and endogenous completeness in finance Dmitry Kramkov (with Silviu Predoiu) Carnegie Mellon University 1 / 19 Bibliography Robert M. Anderson and Roberto C. Raimondo.

More information

On Lead-Lag Estimation

On Lead-Lag Estimation On Lead-Lag Estimation Mathieu Rosenbaum CMAP-École Polytechnique Joint works with Marc Hoffmann, Christian Y. Robert and Nakahiro Yoshida 12 January 2011 Mathieu Rosenbaum On Lead-Lag Estimation 1 Outline

More information

GARCH processes continuous counterparts (Part 2)

GARCH processes continuous counterparts (Part 2) GARCH processes continuous counterparts (Part 2) Alexander Lindner Centre of Mathematical Sciences Technical University of Munich D 85747 Garching Germany lindner@ma.tum.de http://www-m1.ma.tum.de/m4/pers/lindner/

More information

Some Aspects of Universal Portfolio

Some Aspects of Universal Portfolio 1 Some Aspects of Universal Portfolio Tomoyuki Ichiba (UC Santa Barbara) joint work with Marcel Brod (ETH Zurich) Conference on Stochastic Asymptotics & Applications Sixth Western Conference on Mathematical

More information

1 Continuous-time chains, finite state space

1 Continuous-time chains, finite state space Université Paris Diderot 208 Markov chains Exercises 3 Continuous-time chains, finite state space Exercise Consider a continuous-time taking values in {, 2, 3}, with generator 2 2. 2 2 0. Draw the diagramm

More information

Bayesian Regression Linear and Logistic Regression

Bayesian Regression Linear and Logistic Regression When we want more than point estimates Bayesian Regression Linear and Logistic Regression Nicole Beckage Ordinary Least Squares Regression and Lasso Regression return only point estimates But what if we

More information

Multivariate Markov-switching ARMA processes with regularly varying noise

Multivariate Markov-switching ARMA processes with regularly varying noise Multivariate Markov-switching ARMA processes with regularly varying noise Robert Stelzer 23rd June 2006 Abstract The tail behaviour of stationary R d -valued Markov-Switching ARMA processes driven by a

More information

Solutions For Stochastic Process Final Exam

Solutions For Stochastic Process Final Exam Solutions For Stochastic Process Final Exam (a) λ BMW = 20 0% = 2 X BMW Poisson(2) Let N t be the number of BMWs which have passes during [0, t] Then the probability in question is P (N ) = P (N = 0) =

More information

Estimation 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. (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 information

Review. DS GA 1002 Statistical and Mathematical Models. Carlos Fernandez-Granda

Review. DS GA 1002 Statistical and Mathematical Models.   Carlos Fernandez-Granda Review DS GA 1002 Statistical and Mathematical Models http://www.cims.nyu.edu/~cfgranda/pages/dsga1002_fall16 Carlos Fernandez-Granda Probability and statistics Probability: Framework for dealing with

More information

Stochastic process. X, a series of random variables indexed by t

Stochastic process. X, a series of random variables indexed by t Stochastic process X, a series of random variables indexed by t X={X(t), t 0} is a continuous time stochastic process X={X(t), t=0,1, } is a discrete time stochastic process X(t) is the state at time t,

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

Financial Econometrics

Financial Econometrics Financial Econometrics Estimation and Inference Gerald P. Dwyer Trinity College, Dublin January 2013 Who am I? Visiting Professor and BB&T Scholar at Clemson University Federal Reserve Bank of Atlanta

More information

Regular Variation and Extreme Events for Stochastic Processes

Regular Variation and Extreme Events for Stochastic Processes 1 Regular Variation and Extreme Events for Stochastic Processes FILIP LINDSKOG Royal Institute of Technology, Stockholm 2005 based on joint work with Henrik Hult www.math.kth.se/ lindskog 2 Extremes for

More information

Stochastic volatility models: tails and memory

Stochastic volatility models: tails and memory : tails and memory Rafa l Kulik and Philippe Soulier Conference in honour of Prof. Murad Taqqu 19 April 2012 Rafa l Kulik and Philippe Soulier Plan Model assumptions; Limit theorems for partial sums and

More information

Generalized Neyman Pearson optimality of empirical likelihood for testing parameter hypotheses

Generalized Neyman Pearson optimality of empirical likelihood for testing parameter hypotheses Ann Inst Stat Math (2009) 61:773 787 DOI 10.1007/s10463-008-0172-6 Generalized Neyman Pearson optimality of empirical likelihood for testing parameter hypotheses Taisuke Otsu Received: 1 June 2007 / Revised:

More information

Figure 10.1: Recording when the event E occurs

Figure 10.1: Recording when the event E occurs 10 Poisson Processes Let T R be an interval. A family of random variables {X(t) ; t T} is called a continuous time stochastic process. We often consider T = [0, 1] and T = [0, ). As X(t) is a random variable

More information