Determinantal Thinning of Point Processes with Network

Total Page:16

File Type:pdf, Size:1020Kb

Determinantal Thinning of Point Processes with Network Determinantal thinning of point processes with network learning applications B. Błaszczyszyn and H.P. Keeler Inria/ENS, France Abstract—A new type of dependent thinning for point pro- kernel function. But the main obstacle preventing more use of cesses in continuous space is proposed, which leverages the ad- determinantal point processes in R2 (or Rd) is the difficulty vantages of determinantal point processes defined on finite spaces of finding appropriate kernel functions, which need to define and, as such, is particularly amenable to statistical, numerical, and simulation techniques. It gives a new point process that can (integral) operators with eigenvalues in the interval [0, 1]. This serve as a network model exhibiting repulsion. The properties and problem can be largely circumvented when one considers functions of the new point process, such as moment measures, the determinantal point processes defined on spaces with finite car- Laplace functional, the void probabilities, as well as conditional dinality, such as bounded lattices, reducing the mathematical (Palm) characteristics can be estimated accurately by simulating technicalities down to problems of linear algebra. Furthermore, the underlying (non-thinned) point process, which can be taken, for example, to be Poisson. This is in contrast (and preference to) this approach allows the use of non-normalized kernels, which finite Gibbs point processes, which, instead of thinning, require we refer to as L-matrices, to more easily define determinantal weighting the Poisson realizations, involving usually intractable processes. In this setting, Kulesza and Taskar [16] used these normalizing constants. Models based on determinantal point point processes to develop a comprehensive framework for processes are also well suited for statistical (supervised) learning statistical (supervised) learning; also see [15], [17]. techniques, allowing the models to be fitted to observed network patterns with some particular geometric properties. We illustrate We leverage this line of research and define point processes this approach by imitating with determinantal thinning the well- in continuous space using a doubly stochastic approach. First, known Matern´ II hard-core thinning, as well as a soft-core an underlying point process in a bounded subset R ⊂ Rd is thinning depending on nearest-neighbour triangles. These two considered, for which a natural choice is the Poisson point examples demonstrate how the proposed approach can lead to process. Then, the points of a given realization of this process new, statistically optimized, probabilistic transmission scheduling schemes. are considered as a finite, discrete state space on which a Index Terms—dependent thinning, determinantal subset, Palm determinantal process (subset of the realization) is sampled distributions, statistical learning, geometric networks using some kernel that usually depends on the underlying realization. This operation, which can be seen as a dependent I. INTRODUCTION thinning, leads to a new point process existing on bounded regions of Rd and exhibiting more repulsion than the under- Researchers have used point processes on the plane to build lying point process. Conditioned on a given realization of the spatial random models of various wireless network types, but underlying point process, the subset point process inherits all the overwhelming majority of these models relies upon the closed-form expressions available for discrete determinantal Poisson point process [4]. To develop a more realistic model, point processes, thus allowing one to accurately estimate the while still keeping it tractable, we propose a thinning operation characteristics of the new (thinned) point process by simulating using discrete determinantal point processes. the underlying (non-thinned) point process. The statistical Originally called fermion point processes by Macchi [22], learning approach proposed by Kulesza and Taskar [16] can arXiv:1810.08672v2 [cs.LG] 22 Nov 2018 determinantal point processes have attracted considerable at- then be used to fit the kernel of the determinantal thinning to tention in recent years due to their interesting mathemat- various types of observed network models. ical properties [11]. These point processes admit analytic The paper is structured as follows. In Section II we recall approaches to several fundamental characteristics such as the the basics of the determinantal processes in finite spaces; in Laplace functional, the void probabilities and Palm distribu- Section III we introduce the determinantally-thinned point tions [25]. They provide useful statistical models for point processes and some of their characteristics including Palm pattern exhibiting repulsion [3], [19] and, compared to the distributions; in Section IV we present the fitting method well-studied Gibbs point processes [7], have advantages such based on maximum likelihoods; we demonstrate the results as faster simulation methods and more tractable expressions with two illustrative examples in Section V; and in Section VI for likelihoods and moments [18], [19]. This has motivated we discuss network applications. The code for all numerical researchers to use these point processes, when defined on results is available online [12]. the plane R2, as spatial models for base stations in cellular network [10], [20], [21], [24], [27]. II. DETERMINANTAL POINT PROCESSES Determinantal point processes are defined usually via fac- torial moment measures admitting densities in the form of We start by detailing determinantal point processes in a determinants of matrices populated with the values of some discrete setting. A. Finite state space provided all eigenvalues of K are strictly positive, which is P We consider an underlying state space S on which we will equivalent to (Ψ = ∅) > 0. For more details, see, for define a point process (the term carrier space is also used). example, the paper by Borodin and Rains [5, Proposition 1.1] We assume the important simplification that the cardinality of or the book [16] by Kulesza and Taskar. the state space S is finite, that is #(S) < ∞. We consider a III. DETERMINANTALLY-THINNED POINT PROCESSES simple point process Ψ on the state space S, which means that Ψ is a random subset of the state space S, that is Ψ ⊆ S. A We now define a new point process, which builds upon a single realization ψ of this point process Ψ can be interpreted homogeneous Poisson point process Φ with intensity λ > 0 Rd simply as occupied or unoccupied locations in the underlying on a bounded region R⊂ of the n-dimensional Euclidean state space S. space. Given a realization Φ= φ, we consider it as the state space S = φ on which a determinantal point process (subset) B. Definition Ψ ⊂ φ is sampled, resulting in a (typically dependent) thinning For a state space S with finite cardinality m := #(S), a of the realization Φ = φ. More precisely, the points in a discrete point process is a determinantal point process Ψ if for realization φ = {xi}i of a Poisson point process Φ form the all configurations (or subsets) ψ ⊆ S, state space of the finite determinantal point processes Ψ, which is defined via P(Ψ ⊇ ψ) = det(Kψ), (1) P(Ψ ⊇ ψ|Φ= φ) = det(Kψ(φ)), (5) where K is some real symmetric m × m matrix, and Kψ := where K (φ) = [K(φ)] is such that ψ ⊂ φ ⊂ R. Note [K]x,y∈ψ denotes the restriction of K to the entries indexed ψ xi,xj ∈ψ by the elements or points in ψ, that is x, y ∈ ψ. The matrix that Ψ is characterized by intensity measure of the underlying K is called the marginal kernel, and has to be positive semi- Poisson point process Φ and the function K(·) which maps definite. The eigenvalues of K need to be bounded between each (Poisson) realization φ to a semi-definite matrix K(φ) zero and one (with eigenvalues in [0, 1]) having elements indexed by the To simulate or sample a determinantal point process on a points of φ. Rd finite state space, one typically uses an algorithm based on the The point process Ψ is defined on a subset of , but uses eigenvalues and eigenvectors of the matrix K. The number of the discrete approach of determinantal point processes. In other points is given by Bernoulli trials (or biased coin flips) with words, the points of the realization φ are dependently thinned the probabilities of success being equal to the eigenvalues, such that there is repulsion among the points of Ψ. We call this while the joint location of the points is determined by the point process a determinantally-thinned Poisson point process eigenvectors corresponding to the successful trials. Each point or, for brevity, a determintantal Poisson process. is randomly placed one after the other; for further details, The double stochastic construction of determinantally- see [16, Algorithm 1] [19, Algorithm 1] and [28, Algorithm thinned point processes can be compared with the classic 1]. Matern´ hard-core processes (of type I, II and III), which are also constructed through dependent thinning of underlying C. L-ensembles Poisson point processes. For these point processes, there is a In the finite state space setting kernels K can be easily zero probability that any two points are within a certain fixed defined by using the formalism of L-ensembles. Instead of distance of each other. Determinantal thinning of Poisson point finding a K matrix with appropriate eigenvalues, we can processes can provide examples of soft-core process, where work with a family of point processes known as L-ensembles there is a smaller (compared to the Poisson case) probability that are defined through a positive semi-definite matrices L, that any two points are within a certain distance of each other. which is also indexed by the elements of the space S, but We return to this theme later in our results section, where we the eigenvalues of L, though non-negative, do not need to be fit our new point process to a Mat´ern II hard-core process. less than one.
Recommended publications
  • A J-Function for Marked Point Patterns
    AISM (2006) 58: 235–259 DOI 10.1007/s10463-005-0015-7 M.N.M. van Lieshout A J -function for marked point patterns Received: 5 July 2004 / Revised: 7 February 2005 / Published online: 30 May 2006 © The Institute of Statistical Mathematics, Tokyo 2006 Abstract We propose a new summary statistic for marked point patterns. The underlying principle is to compare the distance from a marked point to the near- est other marked point in the pattern to the same distance seen from an arbitrary point in space. Information about the range of interaction can be inferred, and the statistic is well-behaved under random mark allocation. We develop a range of Hanisch style kernel estimators to tackle the problems of exploding tail variance earlier associated with J -function plug-in estimators, and carry out an exploratory analysis of a forestry data set. Keywords Empty space function · J -function · Marked point pattern · Mark correlation function · Nearest neighbour distance distribution function · Product density · Random labelling · Reduced second moment measure · Spatial interaction · Spatial statistics 1 Introduction Marked point patterns are spatial point configurations with a mark attached to each point (Stoyan and Stoyan, 1994). The points could represent the locations in (Euclidean) space of objects, while the marks capture additional information. The latter could be a type label, in which case we also speak of a multivariate point pattern (Cox and Lewis, 1972), a continuous measurement or shape descriptor, or a combination of these. The statistical analysis of such a pattern in general begins by plotting a few summary statistics. Which statistic is used depends on taste and the type of mark.
    [Show full text]
  • A Moment Problem for Random Discrete Measures
    A moment problem for random discrete measures Article Accepted Version Kondratiev, Y. G., Kuna, T. and Lytvynov, E. (2015) A moment problem for random discrete measures. Stochastic Processes and their Applications, 125 (9). pp. 3541-3569. ISSN 0304- 4149 doi: https://doi.org/10.1016/j.spa.2015.03.007 Available at http://centaur.reading.ac.uk/41058/ It is advisable to refer to the publisher’s version if you intend to cite from the work. See Guidance on citing . Published version at: http://dx.doi.org/10.1016/j.spa.2015.03.007 To link to this article DOI: http://dx.doi.org/10.1016/j.spa.2015.03.007 Publisher: Elsevier All outputs in CentAUR are protected by Intellectual Property Rights law, including copyright law. Copyright and IPR is retained by the creators or other copyright holders. Terms and conditions for use of this material are defined in the End User Agreement . www.reading.ac.uk/centaur CentAUR Central Archive at the University of Reading Reading’s research outputs online A moment problem for random discrete measures Yuri G. Kondratiev Fakult¨atf¨urMathematik, Universit¨atBielefeld, Postfach 10 01 31, D-33501 Bielefeld, Germany; NPU, Kyiv, Ukraine e-mail: [email protected] Tobias Kuna University of Reading, Department of Mathematics, Whiteknights, PO Box 220, Read- ing RG6 6AX, U.K. e-mail: [email protected] Eugene Lytvynov Department of Mathematics, Swansea University, Singleton Park, Swansea SA2 8PP, U.K. e-mail: [email protected] Corresponding author: Eugene Lytvynov Department of Mathematics, Swansea University, Singleton Park, Swansea SA2 8PP, U.K.
    [Show full text]
  • Arxiv:2008.08803V1 [Physics.Atom-Ph] 20 Aug 2020
    View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by Caltech Authors - Main Probing Fundamental Symmetries of Deformed Nuclei in Symmetric Top Molecules Phelan Yu∗ and Nicholas R. Hutzlery Division of Physics, Mathematics, and Astronomy, California Institute of Technology, Pasadena, California 91125, USA (Dated: August 21, 2020) Precision measurements of Schiff moments in heavy, deformed nuclei are sensitive probes of beyond Standard Model T;P -violation in the hadronic sector. While the most sensitive limits on Schiff moments to date are set with diamagnetic atoms, polar polyatomic molecules can offer higher sensitivities with unique experimental advantages. In particular, symmetric top molecular ions possess K-doublets of opposite parity with especially small splittings, leading to full polarization at low fields, internal co-magnetometer states useful for rejection of systematic effects, and the ability to perform sensitive searches for T;P -violation using a small number of trapped ions containing heavy 225 + exotic nuclei. We consider the symmetric top cation RaOCH3 as a prototypical and candidate platform for performing sensitive nuclear Schiff measurements and characterize in detail its internal structure using relativistic ab initio methods. The combination of enhancements from a deformed nucleus, large polarizability, and unique molecular structure make this molecule a promising platform to search for fundamental symmetry violation even with a single trapped ion. Searches for permanent electric dipole moments (ThO+, ThF+, AcF, AcO+, AcN, EuO+, EuN, RaO, (EDMs) in atoms and molecules are powerful probes RaF) [26{31] and triatomic molecules (RaOH+, TlOH, of time reversal and parity (T,P-) violating physics ThOH+, TlCN) [27, 32, 33] suitable for Schiff moment posited by beyond Standard Model (BSM) theories [1,2].
    [Show full text]
  • Stability Estimates for Invariant Measures of Diffusion Processes
    Stability estimates for invariant measures of diffusion processes, with applications to stability of moment measures and Stein kernels Max Fathi and Dan Mikulincer October 28, 2020 Abstract We investigate stability of invariant measures of diffusion processes with respect to Lp distances on the coefficients, under an assumption of log- concavity. The method is a variant of a technique introduced by Crippa and De Lellis to study transport equations. As an application, we prove a partial extension of an inequality of Ledoux, Nourdin and Peccati relating transport distances and Stein discrepancies to a non-Gaussian setting via the moment map construction of Stein kernels. 1 Introduction d Let Xt,Yt be stochastic processes in R which satisfy the SDEs, dXt = a(Xt)dt + 2τ(Xt)dBt, dYt = b(Yt)dt + 2σ(Yt)dBt. (1) p p Here Bt is a standard Brownian motion, a, b are vector-valued functions, and σ, τ take values in the cone of symmetric d d positive definite matrices, which we × shall denote by ++. Given X and Y , we shall write the marginal laws of the Sd 0 0 processes as X µ and Y ν . t ∼ t t ∼ t Suppose that, in some sense to be made precise later, a is close to b, and τ is close σ. One can ask whether the measure µt must then be close to νt. Our goal here is to study the quantitative regime of this problem. The method used here arXiv:2010.14178v1 [math.PR] 27 Oct 2020 is an adaptation of a technique developed by Crippa and De Lellis for transport equations.
    [Show full text]
  • A Moment Problem for Pseudo-Positive Definite Functionals
    A moment problem for pseudo-positive definite functionals Ognyan Kounchev and Hermann Render October 31, 2018 Abstract A moment problem is presented for a class of signed measures which are termed pseudo-positive. Our main result says that for every pseudo- positive definite functional (subject to some reasonable restrictions) there exists a representing pseudo-positive measure. The second main result is a characterization of determinacy in the class of equivalent pseudo-positive representation measures. Finally the corresponding truncated moment problem is discussed. Key words: Multidimensional moment problem, pseudopositive measures, spherical harmonics, multidimensional numerical integration. MSC 2000 classification: 43A32, 47A57, 65D32 1 Introduction Let C [x1, ..., xd] denote the space of all polynomials in d variables with complex coefficients and let T : C [x1, ..., xd] C be a linear functional. The multivariate moment problem asks for conditions→ on the functional T such that there exists a non-negative measure µ on Rd with T (P )= P (x) dµ (x) (1) Rd Z arXiv:0802.0023v1 [math.FA] 31 Jan 2008 for all P C [x1, ..., xd] . It is well known that positive definiteness of the func- tional T is∈ a necessary condition which means that ∗ T (P P ) 0 for all P C [x , ..., xd]; ≥ ∈ 1 here P ∗ is the polynomial whose coefficients are the complex conjugates of the coefficients of P. By a theorem of Haviland, a necessary and sufficient condition for the existence of a non-negative measure µ satisfying (1) is the positivity of the functional T , i.e. P (x) 0 for all x Rd implies T (P ) 0 for all ≥ ∈ ≥ P C [x1, ..., xd], cf.
    [Show full text]
  • A Nonparametric Measure of Spatial Interaction in Point Patterns
    344 Statistica Neerlandica (1996) Vol. 50, nr. 3, pp. 344-361 A nonparametric measure of spatial interaction in point patterns M. N. M. van Lieshout~ Department of Statistics, University of Warwick, Coventry CV4 7AL, United Kingdom A. J. Baddeley Department of Mathematics, University of Western Australia, Nedlands WA 6907, Australia, and Department of Mathematics and Computer Science, University of Leiden, The Netherlands In memory of Philip Holgate The strength and range of interpoint interactions in a spatial point process can be quantified by the function J = (1 - G) I (1 - F), where G is the nearest-neighbour distance distribution function and F the empty space function of the process. J (r) is identically equal to 1 for a Poisson process; values of J(r) smaller or larger than 1 indicate clustering or regularity, respectively. We show that, for a large class of point pro­ cesses, J (r) is constant for distances r greater than the range of spatial interaction. Hence both the range and type of interaction can be inferred from J without parametric model assumptions. It is also possible to evaluate J (r) explicitly for many point process models, so that J is also useful for parameter estimation. Various properties are derived, including the fact that the J function of the superposition of independent point processes is a weighted mean of the J functions of the individual processes. Estimators of J can be constructed from standard estimators of F and G. We compute estimates of J for several standard point pattern datasets and implement a Monte Carlo test for complete spatial random­ ness.
    [Show full text]
  • Affine Moments of a Random Vector
    1 Affine moments of a random vector Erwin Lutwak, Songjun Lv, Deane Yang, and Gaoyong Zhang Abstract—An affine invariant p-th moment measure is defined explicit integral formula for this moment is derived. We then for a random vector and used to prove sharp moment-entropy establish sharp information inequalities giving a sharp lower inequalities that are more general and stronger than standard bound of Renyi´ entropy in terms of the affine moment. These moment-entropy inequalities. new affine moment-entropy inequalities imply the moment- Index Terms—Moment, affine moment, information measure, entropy inequalities obtained in [32], [35]. information theory, entropy, Renyi´ entropy, Shannon entropy, In [26] an approach similar to the one taken here is used Shannon theory, entropy inequality, Gaussian, generalized Gaus- sian. to define an affine invariant version of Fisher information, and a corresponding sharp Fisher information inequality is established. I. INTRODUCTION It is worth noting that the affine moment measure introduced The moment-entropy inequality states that a continuous ran- here, as well as the affine Fisher information studied in [26], dom variable with given second moment has maximal Shannon is closely related to analogous affine invariant generalizations entropy if and only if it is a Gaussian (see, for example, [15], of the surface area of a convex body (see, for example, [9], Theorem 9.6.5). The Fisher information inequality states that [27], [30]). In fact, the results presented here are part of an a continuous random variable with given Fisher information ongoing effort by the authors to explore connections between has minimal Shannon entropy if and only if it is a Gaussian information theory and geometry.
    [Show full text]
  • Mathematical Statistics in Physics and Astronomy
    Mathematical statistics in physics and astronomy Radu Stoica Universit´eLille - Laboratoire Paul Painlev´e Observatoire de Paris - Institut de M´ecanique C´eleste et Calcul d Eph´em´erides [email protected] Tartu, September, 2014 About me ◮ associate professor at Universit´eLille ◮ working together with the Toravere Observatory since 2003 ◮ mail : [email protected] ◮ web page : http://math.univ-lille1.fr/ stoica/ ◮ office : Toravere Observatory (during September) ◮ phone : + 33 6 20 06 29 30 Course structure ◮ 24 hours = 4 weeks 6 hours = 4 weeks 2 days times 3 × × hours ◮ mark : project + written/oral exam (last day of the course - you need a mark of 50% to validate it) Bibliogaphical references Baddeley, A. J. (2010) Analysing spatial point patterns in ’R’. • Workshop Notes, Version 4.1. Diggle, P. J. (2003) Statistical analysis of spatial point patterns. • Arnold Publishers. van Lieshout, M. N. M. (2000) Markov point processes and their • applications. Imperial College Press. Martinez, V. J., Saar, E. (2002) Statistics of the galaxy • distribution. Chapman and Hall. Møller, J., Waagepetersen, R. P. (2004) Statistical inference and • simulation for spatial point processes. Chapman and Hall/CRC.V Chiu, S. N., Stoyan, D., Kendall, W. S., Mecke, J. (2013) • Stochastic geometry and its applications. John Wiley and Sons. Aknowledgements A. J. Baddeley, G. Castellan, J.-F. Coeurjolly, X. Descombes, P. Gregori, C. Lantu´ejoul, M. N. M. van Liehsout, V. Martinez, J. Mateu, J. Møller, Z. Pawlas, F. Rodriguez, E. Saar, D.
    [Show full text]
  • NMTP438, Topic 4: Binomial, Poisson and Cox Point Process
    NMTP438, topic 4: binomial, Poisson and Cox point process 1. Show that the mixed binomial point process with the Poisson distribution (with parameter λ) of the ν(·) number of points N is a Poisson process with the intensity measure λ ν(B) . 2. Let Φ be a Poisson point process with the intensity measure Λ and B 2 B be a given Borel set. Show that ΦjB is a Poisson point process and determine its intensity measure. 3. Consider two independent Poisson point processes Φ1 and Φ2 with the intensity measures Λ1 and Λ2. Show that Φ = Φ1 + Φ2 is a Poisson process and determine its intensity measure. 4. Let Φ be a Poisson point process with the intensity measure Λ. Determine the covariance cov(Φ(B1); Φ(B2)) for B1;B2 2 B. 5. Let Φ be a binomial point process with n points in B and the measure ν. Determine the covariance cov(Φ(B1); Φ(B2)) for B1;B2 2 B. 6. Let Φ be a mixed Poisson point process with the driving measure Y · Λ, where Y is a non-negative random variable and Λ is a locally finite diffuse measure. Determine the covariance cov(Φ(B1); Φ(B2)) for B1;B2 2 B0 and show that it is non-negative. 7. Determine the second-order factorial moment measure of a binomial point process. 8. Determine the Laplace transform of a binomial point process. 9. Dispersion of a random variable Φ(B) is defined as var Φ(B) D(Φ(B)) = ;B 2 B0: E Φ(B) Show that a) for a Poisson process D(Φ(B)) = 1, b) a binomial process is underdispersed, i.e.
    [Show full text]
  • Arxiv:1811.08174V1 [Math.PR] 20 Nov 2018 Rcse Hc R Qiain Ne Oednmc.Tedifferen the Dynamics
    ERGODIC POISSON SPLITTINGS ELISE´ JANVRESSE, EMMANUEL ROY AND THIERRY DE LA RUE Abstract. In this paper we study splittings of a Poisson point process which are equivariant under a conservative transformation. We show that, if the Cartesian powers of this transformation are all ergodic, the only ergodic split- ting is the obvious one, that is, a collection of independent Poisson processes. We apply this result to the case of a marked Poisson process: under the same hypothesis, the marks are necessarily independent of the point process and i.i.d. Under additional assumptions on the transformation, a further applica- tion is derived, giving a full description of the structure of a random measure invariant under the action of the transformation. 1. Introduction Thinning and splitting are classical operations when studying point processes. Thinning consists in removing points according to some rule, whereas the related notion of splitting means decomposing the point process as the sum of several other point processes. It is well known that thinning a Poisson point process by choosing to remove points according to independent coin tosses yields a new Poisson process of lower (but proportional) intensity. Moreover, this procedure gives rise to a splitting of the original Poisson process into a sum of two independent Poisson processes. In recent years, some new results on thinnings of Poisson process have emerged. In particular, it is shown in [2] that it is possible to deterministically choose points from a homogeneous Poisson point process on R to get another homogeneous Pois- son process of lower intensity. Moreover, it is possible to proceed in a translation equivariant way.
    [Show full text]
  • Spatial Point Process Theory
    16 Spatial Point Process Theory Marie-Colette van Lieshout CONTENTS 16.1 Spatial Point Process Theory ..........................................................................................263 16.1.1 Introduction ........................................................................................................263 16.1.2 Characterization of Point Process Distributions............................................266 16.1.3 Campbell and Moment Measures....................................................................267 16.1.4 Reduced and Higher-Order Campbell Measures..........................................270 16.1.5 Palm Theory and Conditioning .......................................................................273 16.1.6 Finite Point Processes ........................................................................................278 16.1.7 Gibbs Measures by Local Specification...........................................................280 References.....................................................................................................................................281 16.1 Spatial Point Process Theory 16.1.1 Introduction A spatial point process is a stochastic process each of whose realizations consists of a finite or countably infinite set of points in the plane. This chapter sets out the mathematical theory of such processes. After some motivation, several ways to characterize the distribution of a point process are described in Section 16.1.2. The important notions of Campbell and moment measures are introduced
    [Show full text]
  • Lectures on the Poisson Process
    Lectures on the Poisson Process Gunter¨ Last and Mathew Penrose, Version from 21 August 2017, Commercial reproduction prohibited, to be published as IMS Textbook by Cambridge University Press, c Gunter¨ Last and Mathew Penrose Gunter¨ Last Institut f¨urStochastik, Karlsruhe Institute of Technology, Englerstraße 2, D-76128 Karlsruhe, Germany, [email protected] Mathew Penrose Department of Mathematical Sciences, University of Bath, Bath BA2 7AY, United Kingdom, [email protected] Contents Preface page x List of Symbols xiii 1 Poisson and Other Discrete Distributions 1 1.1 The Poisson Distribution1 1.2 Relationships Between Poisson and Binomial Distributions3 1.3 The Poisson Limit Theorem4 1.4 The Negative Binomial Distribution5 1.5 Exercises7 2 Point Processes 9 2.1 Fundamentals9 2.2 Campbell’s Formula 12 2.3 Distribution of a Point Process 14 2.4 Point Processes on Metric Spaces 16 2.5 Exercises 18 3 Poisson Processes 19 3.1 Definition of the Poisson Process 19 3.2 Existence of Poisson Processes 20 3.3 Laplace Functional of the Poisson Process 23 3.4 Exercises 24 4 The Mecke Equation and Factorial Measures 26 4.1 The Mecke Equation 26 4.2 Factorial Measures and the Multivariate Mecke Equation 28 4.3 Janossy Measures 32 4.4 Factorial Moment Measures 34 4.5 Exercises 36 v vi Contents 5 Mappings, Markings and Thinnings 38 5.1 Mappings and Restrictions 38 5.2 The Marking Theorem 39 5.3 Thinnings 42 5.4 Exercises 44 6 Characterisations of the Poisson Process 46 6.1 Borel Spaces 46 6.2 Simple Point Processes 49 6.3 Renyi’s´ Theorem 50
    [Show full text]