DOCSLIB.ORG

Complete Dissertation

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Complete Dissertation VU Research Portal Time-Varying Parameter Models for Discrete Valued Time Series Lit, R. 2016 document version Publisher's PDF, also known as Version of record Link to publication in VU Research Portal citation for published version (APA) Lit, R. (2016). Time-Varying Parameter Models for Discrete Valued Time Series. Tinbergen Institute. General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal ? Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. E-mail address: [email protected] Download date: 05. Oct. 2021 TIME-VARYING PARAMETER MODELS FOR DISCRETE VALUED TIME SERIES ISBN 978 90 5170 755 7 Cover design: Crasborn Graphic Designers bno, Valkenburg a.d. Geul Image: Wiki Commons (template) and Marcin Zamojski (design) This book is no. 642 of the Tinbergen Institute Research Series, established through coopera- tion between Rozenberg Publishers and the Tinbergen Institute. A list of books which already appeared in the series can be found in the back. VRIJE UNIVERSITEIT TIME-VARYING PARAMETER MODELS FOR DISCRETE VALUED TIME SERIES ACADEMISCH PROEFSCHRIFT ter verkrijging van de graad Doctor aan de Vrije Universiteit Amsterdam, op gezag van de rector magnificus prof.dr. V. Subramaniam, in het openbaar te verdedigen ten overstaan van de promotiecommissie van de Faculteit der Economische Wetenschappen en Bedrijfskunde op dinsdag 9 februari 2016 om 11.45 uur in de aula van de universiteit, De Boelelaan 1105 door Rutger Lit geboren te Alkmaar promotoren: prof.dr. A. Lucas prof.dr. S.J. Koopman Acknowledgements The last two and a half years went by at a fast pace. It was a period of time that I greatly enjoyed and the end product, this dissertation, is something that I am proud of. I could not have done this research alone. I am especially grateful to my supervisors Siem Jan Koopman and Andr´e Lucas who were essential for the realisation of this dissertation. Their expert advice on numerous topics in combination with frequent meetings, despite their busy schedules, were very important and much appreciated. I look forward to continuing our collaboration. From all my colleagues at the finance and econometrics department at the VU Am- sterdam, a special thanks goes out to Istv´an Barra. The originality and quality of our research discussions were very helpful to me and very much appreciated. Another thanks goes out to Marcin Zamojski for his advice on several econometric and finance topics as well as making my digital life easier and more secure. I would also like to thank Marco Bazzi, Falk Br¨auning, Artem Duplinskiy, Wenqian Huang (Ä)), Anne Opschoor, Kristian Støre, Andries van Vlodrop, and Marius Zoican for the discussions and laughs we shared. I am grateful to the Dutch National Science Foundation (NWO) and the Tinbergen Institute for the financial support. Last but not least I thank my parents for never doubting the research path that I have chosen. Amsterdam, November 2015 Contents Acknowledgements v 1 Introduction 1 1.1 Introduction for a general audience ......................... 1 1.2Econometricmethodologies............................. 2 1.2.1 Non-Gaussianstatespacemodels...................... 2 1.2.2 Score-drivenmodels.............................. 4 1.3Contributions..................................... 4 1.3.1 Chapter2................................... 4 1.3.2 Chapter3................................... 5 1.3.3 Chapter4................................... 6 1.3.4 Chapter5................................... 7 2 A Dynamic Bivariate Poisson Model for Analysing Football Match Results 9 2.1Introduction...................................... 9 2.2Thestatisticalmodellingframework......................... 12 2.2.1 BivariatePoissonmodel........................... 12 2.2.2 Dynamic specification for goal scoring intensities .............. 13 2.2.3 Someextensionsofthebasicmodel..................... 14 2.2.4 General state space representation of the model .............. 16 2.2.5 Evaluation of likelihood function and estimation .............. 17 2.3 Empirical application ................................. 19 2.3.1 Data description ............................... 19 2.3.2 Detailsofthebasicmodel.......................... 20 2.3.3 Parameterestimates............................. 21 2.3.4 Signal estimates of strengths of attack and defence ............ 22 2.3.5 Model evaluations: in-sample and out-of-sample .............. 24 2.4 Out-of-sample performance in a betting strategy .................. 28 2.5Conclusions...................................... 32 Appendices......................................... 33 A Likelihoodevaluation............................. 33 B Constructionofapproximatingmodel.................... 34 C Thederivativesforthemodelobservationdensity............. 36 D Computational details ............................ 37 E Tablesandfigures............................... 40 3 Intraday Stochastic Volatility in Discrete Price Changes 49 3.1Introduction...................................... 49 3.2ThedynamicSkellammodel............................. 52 3.2.1 The Skellam distribution ........................... 52 3.2.2 The modified Skellam distribution ...................... 52 3.2.3 TheSkellammodelwithdynamicmeanandvariance........... 55 3.3Analysisofhigh-frequencySkellampricechanges................. 57 3.3.1 Data...................................... 57 3.3.2 Dynamic Skellam with Intraday Stochastic Volatility ........... 58 3.3.3 Parameterestimationresults......................... 60 3.3.4 Signal extraction ............................... 62 3.3.5 Goodness-of-fit ................................ 62 3.3.6 Diagnosticchecking.............................. 64 3.3.7 Forecastingstudy............................... 66 3.4Conclusions...................................... 69 Appendices......................................... 70 A Modified Skellam distribution of type I ................... 70 B Moments of the MSKII(i, j, k) distribution ................. 70 C Simulationstudy............................... 71 D Numerically accelerated importance sampling ............... 73 E Intradaily time series of price changes in 2012 ............... 77 4 A Skellam Model for Analysing the Differences in Count Data 79 4.1Introduction...................................... 79 4.2ThedynamicSkellammodel............................. 80 4.2.1 Skellam distribution ............................. 80 4.2.2 Dynamicspecificationofintensities..................... 81 4.3Analysingfootballscores............................... 82 4.3.1 Estimationresults............................... 84 4.3.2 Modelextensions............................... 85 4.3.3 Signal extraction ............................... 88 4.4Conclusions...................................... 90 viii 5 Dynamic Discrete Copula Models for High Frequency Stock Price Changes 93 5.1Introduction...................................... 93 5.2 Score-driven dynamic discrete copula model .................... 96 5.3Simulationstudy...................................100 5.3.1 Estimatingparameterswhenmodeliscorrectlyspecified.........100 5.3.2 Estimatingtime-varyingpathswhenmodelismisspecified........101 5.4Dependencebetweendiscretepricechanges.....................103 5.4.1 Data description ...............................104 5.4.2 Missingvalues.................................105 5.4.3 Copula selection ................................106 5.4.4 Fullyearresults................................107 5.4.5 Comparison with intraday spline ......................110 5.5Conclusions......................................111 Appendices.........................................114 A Derivationofthescorevector........................114 B Tablesandfigures...............................115 Bibliography 119 Summary 127 ix Chapter 1 Introduction The introduction of this dissertation is organised as follows. Section 1.1 explains the title of this dissertation to a general audience and uses a football example as a lively illustration of the use of time-varying parameter models. Here, the use of mathematical notation is avoided as much as possible. Starting from Section 1.2, the focus is on the more experienced reader in the field of econometrics and statistics. 1.1 Introduction for a general audience Suppose in the very unlikely event that a football aficionado with knowledge of econo- metrics, statistics, and finance is interested in predicting the outcome of the next football match. This dissertation, with the title ‘time-varying parameter models for discrete valued time series’, can assist with these predictions. Let us start by getting a good understanding of what a time series is. Here, I consider a time series as a set of observations taken at (possibly unequally spaced) intervals over time. For example, we can observe the number of goals scored by a football team over a period of time (for example each week for a period of five years). It is crucial that the order of the observations
Load more

Recommended publications
  • Arxiv:2004.02679V2 [Math.OA] 17 Jul 2020
    THE FREE TANGENT LAW WIKTOR EJSMONT AND FRANZ LEHNER Abstract. Nevanlinna-Herglotz functions play a fundamental role for the study of infinitely divisible distributions in free probability [11]. In the present paper we study the role of the tangent function, which is a fundamental Herglotz-Nevanlinna function [28, 23, 54], and related functions in free probability. To be specific, we show that the function tan z 1 ´ x tan z of Carlitz and Scoville [17, (1.6)] describes the limit distribution of sums of free commutators and anticommutators and thus the free cumulants are given by the Euler zigzag numbers. 1. Introduction Nevanlinna or Herglotz functions are functions analytic in the upper half plane having non- negative imaginary part. This class has been thoroughly studied during the last century and has proven very useful in many applications. One of the fundamental examples of Nevanlinna functions is the tangent function, see [6, 28, 23, 54]. On the other hand it was shown by by Bercovici and Voiculescu [11] that Nevanlinna functions characterize freely infinitely divisible distributions. Such distributions naturally appear in free limit theorems and in the present paper we show that the tangent function appears in a limit theorem for weighted sums of free commutators and anticommutators. More precisely, the family of functions tan z 1 x tan z ´ arises, which was studied by Carlitz and Scoville [17, (1.6)] in connection with the combinorics of tangent numbers; in particular we recover the tangent function for x 0. In recent years a number of papers have investigated limit theorems for“ the free convolution of probability measures defined by Voiculescu [58, 59, 56].
    [Show full text]
  • A Guide on Probability Distributions
    powered project A guide on probability distributions R-forge distributions Core Team University Year 2008-2009 LATEXpowered Mac OS' TeXShop edited Contents Introduction 4 I Discrete distributions 6 1 Classic discrete distribution 7 2 Not so-common discrete distribution 27 II Continuous distributions 34 3 Finite support distribution 35 4 The Gaussian family 47 5 Exponential distribution and its extensions 56 6 Chi-squared's ditribution and related extensions 75 7 Student and related distributions 84 8 Pareto family 88 9 Logistic ditribution and related extensions 108 10 Extrem Value Theory distributions 111 3 4 CONTENTS III Multivariate and generalized distributions 116 11 Generalization of common distributions 117 12 Multivariate distributions 132 13 Misc 134 Conclusion 135 Bibliography 135 A Mathematical tools 138 Introduction This guide is intended to provide a quite exhaustive (at least as I can) view on probability distri- butions. It is constructed in chapters of distribution family with a section for each distribution. Each section focuses on the tryptic: definition - estimation - application. Ultimate bibles for probability distributions are Wimmer & Altmann (1999) which lists 750 univariate discrete distributions and Johnson et al. (1994) which details continuous distributions. In the appendix, we recall the basics of probability distributions as well as \common" mathe- matical functions, cf. section A.2. And for all distribution, we use the following notations • X a random variable following a given distribution, • x a realization of this random variable, • f the density function (if it exists), • F the (cumulative) distribution function, • P (X = k) the mass probability function in k, • M the moment generating function (if it exists), • G the probability generating function (if it exists), • φ the characteristic function (if it exists), Finally all graphics are done the open source statistical software R and its numerous packages available on the Comprehensive R Archive Network (CRAN∗).
    [Show full text]
  • An Alternative Discrete Skew Logistic Distribution
    An Alternative Discrete Skew Logistic Distribution Deepesh Bhati1*, Subrata Chakraborty2, and Snober Gowhar Lateef1 1Department of Statistics, Central University of Rajasthan, 2Department of Statistics, Dibrugarh University, Assam *Corresponding Author: [email protected] April 7, 2016 Abstract In this paper, an alternative Discrete skew Logistic distribution is proposed, which is derived by using the general approach of discretizing a continuous distribution while retaining its survival function. The properties of the distribution are explored and it is compared to a discrete distribution defined on integers recently proposed in the literature. The estimation of its parameters are discussed, with particular focus on the maximum likelihood method and the method of proportion, which is particularly suitable for such a discrete model. A Monte Carlo simulation study is carried out to assess the statistical properties of these inferential techniques. Application of the proposed model to a real life data is given as well. 1 Introduction Azzalini A.(1985) and many researchers introduced different skew distributions like skew-Cauchy distribu- tion(Arnold B.C and Beaver R.J(2000)), Skew-Logistic distribution (Wahed and Ali (2001)), Skew Student's t distribution(Jones M.C. et. al(2003)). Lane(2004) fitted the existing skew distributions to insurance claims data. Azzalini A.(1985), Wahed and Ali (2001) developed Skew Logistic distribution by taking f(x) to be Logistic density function and G(x) as its CDF of standard Logistic distribution, respectively and obtained arXiv:1604.01541v1 [stat.ME] 6 Apr 2016 the probability density function(pdf) as 2e−x f(x; λ) = ; −∞ < x < 1; −∞ < λ < 1 (1 + e−x)2(1 + e−λx) They have numerically studied cdf, moments, median, mode and other properties of this distribution.
    [Show full text]
  • Skellam Type Processes of Order K and Beyond
    Skellam Type Processes of Order K and Beyond Neha Guptaa, Arun Kumara, Nikolai Leonenkob a Department of Mathematics, Indian Institute of Technology Ropar, Rupnagar, Punjab - 140001, India bCardiff School of Mathematics, Cardiff University, Senghennydd Road, Cardiff, CF24 4AG, UK Abstract In this article, we introduce Skellam process of order k and its running average. We also discuss the time-changed Skellam process of order k. In particular we discuss space-fractional Skellam process and tempered space-fractional Skellam process via time changes in Poisson process by independent stable subordinator and tempered stable subordinator, respectively. We derive the marginal proba- bilities, L´evy measures, governing difference-differential equations of the introduced processes. Our results generalize Skellam process and running average of Poisson process in several directions. Key words: Skellam process, subordination, L´evy measure, Poisson process of order k, running average. 1 Introduction Skellam distribution is obtained by taking the difference between two independent Poisson distributed random variables which was introduced for the case of different intensities λ1, λ2 by (see [1]) and for equal means in [2]. For large values of λ1 + λ2, the distribution can be approximated by the normal distribution and if λ2 is very close to 0, then the distribution tends to a Poisson distribution with intensity λ1. Similarly, if λ1 tends to 0, the distribution tends to a Poisson distribution with non- positive integer values. The Skellam random variable is infinitely divisible since it is the difference of two infinitely divisible random variables (see Prop. 2.1 in [3]). Therefore, one can define a continuous time L´evy process for Skellam distribution which is called Skellam process.
    [Show full text]
  • On the Bivariate Skellam Distribution Jan Bulla, Christophe Chesneau, Maher Kachour
    On the bivariate Skellam distribution Jan Bulla, Christophe Chesneau, Maher Kachour To cite this version: Jan Bulla, Christophe Chesneau, Maher Kachour. On the bivariate Skellam distribution. 2012. hal- 00744355v1 HAL Id: hal-00744355 https://hal.archives-ouvertes.fr/hal-00744355v1 Preprint submitted on 22 Oct 2012 (v1), last revised 23 Oct 2013 (v2) HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Noname manuscript No. (will be inserted by the editor) On the bivariate Skellam distribution Jan Bulla · Christophe Chesneau · Maher Kachour Submitted to Journal of Multivariate Analysis: 22 October 2012 Abstract In this paper, we introduce a new distribution on Z2, which can be viewed as a natural bivariate extension of the Skellam distribution. The main feature of this distribution a possible dependence of the univariate components, both following univariate Skellam distributions. We explore various properties of the distribution and investigate the estimation of the unknown parameters via the method of moments and maximum likelihood. In the experimental section, we illustrate our theory. First, we compare the performance of the estimators by means of a simulation study. In the second part, we present and an application to a real data set and show how an improved fit can be achieved by estimating mixture distributions.
    [Show full text]
  • A CURE for Noisy Magnetic Resonance Images: Chi-Square Unbiased Risk Estimation
    SUBMITTED MANUSCRIPT 1 A CURE for Noisy Magnetic Resonance Images: Chi-Square Unbiased Risk Estimation Florian Luisier, Thierry Blu, and Patrick J. Wolfe Abstract In this article we derive an unbiased expression for the expected mean-squared error associated with continuously differentiable estimators of the noncentrality parameter of a chi- square random variable. We then consider the task of denoising squared-magnitude magnetic resonance image data, which are well modeled as independent noncentral chi-square random variables on two degrees of freedom. We consider two broad classes of linearly parameterized shrinkage estimators that can be optimized using our risk estimate, one in the general context of undecimated filterbank transforms, and another in the specific case of the unnormalized Haar wavelet transform. The resultant algorithms are computationally tractable and improve upon state-of-the-art methods for both simulated and actual magnetic resonance image data. EDICS: TEC-RST (primary); COI-MRI, SMR-SMD, TEC-MRS (secondary) Florian Luisier and Patrick J. Wolfe are with the Statistics and Information Sciences Laboratory, Harvard University, Cambridge, MA 02138, USA (email: fl[email protected], [email protected]) Thierry Blu is with the Department of Electronic Engineering, The Chinese University of Hong Kong, Shatin, N.T., Hong Kong (e-mail: [email protected]). arXiv:1106.2848v1 [stat.AP] 15 Jun 2011 This work was supported by the Swiss National Science Foundation Fellowship BELP2-133245, the US National Science Foundation Grant DMS-0652743, and the General Research Fund CUHK410209 from the Hong Kong Research Grant Council. August 20, 2018 DRAFT SUBMITTED MANUSCRIPT 2 I.
    [Show full text]
  • Handbook on Probability Distributions
    R powered R-forge project Handbook on probability distributions R-forge distributions Core Team University Year 2009-2010 LATEXpowered Mac OS' TeXShop edited Contents Introduction 4 I Discrete distributions 6 1 Classic discrete distribution 7 2 Not so-common discrete distribution 27 II Continuous distributions 34 3 Finite support distribution 35 4 The Gaussian family 47 5 Exponential distribution and its extensions 56 6 Chi-squared's ditribution and related extensions 75 7 Student and related distributions 84 8 Pareto family 88 9 Logistic distribution and related extensions 108 10 Extrem Value Theory distributions 111 3 4 CONTENTS III Multivariate and generalized distributions 116 11 Generalization of common distributions 117 12 Multivariate distributions 133 13 Misc 135 Conclusion 137 Bibliography 137 A Mathematical tools 141 Introduction This guide is intended to provide a quite exhaustive (at least as I can) view on probability distri- butions. It is constructed in chapters of distribution family with a section for each distribution. Each section focuses on the tryptic: definition - estimation - application. Ultimate bibles for probability distributions are Wimmer & Altmann (1999) which lists 750 univariate discrete distributions and Johnson et al. (1994) which details continuous distributions. In the appendix, we recall the basics of probability distributions as well as \common" mathe- matical functions, cf. section A.2. And for all distribution, we use the following notations • X a random variable following a given distribution, • x a realization of this random variable, • f the density function (if it exists), • F the (cumulative) distribution function, • P (X = k) the mass probability function in k, • M the moment generating function (if it exists), • G the probability generating function (if it exists), • φ the characteristic function (if it exists), Finally all graphics are done the open source statistical software R and its numerous packages available on the Comprehensive R Archive Network (CRAN∗).
    [Show full text]
  • Package 'Extradistr'
    Package ‘extraDistr’ September 7, 2020 Type Package Title Additional Univariate and Multivariate Distributions Version 1.9.1 Date 2020-08-20 Author Tymoteusz Wolodzko Maintainer Tymoteusz Wolodzko <[email protected]> Description Density, distribution function, quantile function and random generation for a number of univariate and multivariate distributions. This package implements the following distributions: Bernoulli, beta-binomial, beta-negative binomial, beta prime, Bhattacharjee, Birnbaum-Saunders, bivariate normal, bivariate Poisson, categorical, Dirichlet, Dirichlet-multinomial, discrete gamma, discrete Laplace, discrete normal, discrete uniform, discrete Weibull, Frechet, gamma-Poisson, generalized extreme value, Gompertz, generalized Pareto, Gumbel, half-Cauchy, half-normal, half-t, Huber density, inverse chi-squared, inverse-gamma, Kumaraswamy, Laplace, location-scale t, logarithmic, Lomax, multivariate hypergeometric, multinomial, negative hypergeometric, non-standard beta, normal mixture, Poisson mixture, Pareto, power, reparametrized beta, Rayleigh, shifted Gompertz, Skellam, slash, triangular, truncated binomial, truncated normal, truncated Poisson, Tukey lambda, Wald, zero-inflated binomial, zero-inflated negative binomial, zero-inflated Poisson. License GPL-2 URL https://github.com/twolodzko/extraDistr BugReports https://github.com/twolodzko/extraDistr/issues Encoding UTF-8 LazyData TRUE Depends R (>= 3.1.0) LinkingTo Rcpp 1 2 R topics documented: Imports Rcpp Suggests testthat, LaplacesDemon, VGAM, evd, hoa,
    [Show full text]
  • On the Poisson Difference Distribution Inference and Applications 19
    BULLETIN of the Bull. Malays. Math. Sci. Soc. (2) 33(1) (2010), 17–45 Malaysian Mathematical Sciences Society http://math.usm.my/bulletin On The Poisson Difference Distribution Inference and Applications 1Abdulhamid A. Alzaid and 2Maha A. Omair 1;2Department of Statistics and Operations Research, College of Sciences, King Saud University, BOX 2455 Riyadh 11451, Kingdom of Saudi Arabia [email protected], [email protected] Abstract. The distribution of the difference between two independent Poisson random variables involves the modified Bessel function of the first kind. Using properties of this function, maximum likelihood estimates of the parameters of the Poisson difference were derived. Asymptotic distribution property of the maximum likelihood estimates is discussed. Maximum likelihood estimates were compared with the moment estimates in a Monte Carlo study. Hypothesis test- ing using likelihood ratio tests was considered. Some new formulas concerning the modified Bessel function of the first kind were provided. Alternative formu- las for the probability mass function of the Poisson difference (PD) distribution are introduced. Finally, two new applications for the PD distribution are pre- sented. The first is from the Saudi stock exchange (TASI) and the second is from Dallah hospital. 2000 Mathematics Subject Classification: Primary 60E05; Secondary 62F86, 46N30 Key words and phrases: Poisson difference distribution, Skellam distribution, Bessel function, regularized hypergeometric function, maximum likelihood esti- mate, likelihood ratio test. 1. Introduction The distribution of the difference between two independent Poisson random variables was derived by Irwin [5] for the case of equal parameters. Skellam [13] and Prekopa [11] discussed the case of unequal parameters.
    [Show full text]
  • Ebookdistributions.Pdf
    DOWNLOAD YOUR FREE MODELRISK TRIAL Adapted from Risk Analysis: a quantitative guide by David Vose. Published by John Wiley and Sons (2008). All Rights Reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except under the terms of the Copyright, Designs and Patents Act 1988 or under the terms of a licence issued by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London W1T 4LP, UK, without the permission in writing of the Publisher. If you notice any errors or omissions, please contact [email protected] Referencing this document Please use the following reference, following the Harvard system of referencing: Van Hauwermeiren M, Vose D and Vanden Bossche S (2012). A Compendium of Distributions (second edition). [ebook]. Vose Software, Ghent, Belgium. Available from www.vosesoftware.com . Accessed dd/mm/yy. © Vose Software BVBA (2012) www.vosesoftware.com Updated 17 January, 2012. Page 2 Table of Contents Introduction .................................................................................................................... 7 DISCRETE AND CONTINUOUS DISTRIBUTIONS.......................................................................... 7 Discrete Distributions .............................................................................................. 7 Continuous Distributions ........................................................................................
    [Show full text]
  • Arxiv:1710.02036V4 [Cs.CR] 1 Dec 2017 ‹ Hsi H Ulvrino 3] H Eerhwsspotdb Supported Was Research the [38]
    Computational Differential Privacy from Lattice-based Cryptography ‹ Filipp Valovich and Francesco Ald`a Horst G¨ortz Institute for IT Security Faculty of Mathematics Ruhr-Universit¨at Bochum, Universit¨atsstrasse 150, 44801 Bochum, Germany {filipp.valovich,francesco.alda}@rub.de Abstract. The emerging technologies for large scale data analysis raise new challenges to the security and privacy of sensitive user data. In this work we investigate the problem of private statistical analysis of time-series data in the distributed and semi-honest setting. In particu- lar, we study some properties of Private Stream Aggregation (PSA), first introduced by Shi et al. 2011. This is a computationally secure proto- col for the collection and aggregation of data in a distributed network and has a very small communication cost. In the non-adaptive query model, a secure PSA scheme can be built upon any key-homomorphic weak pseudo-random function as shown by Valovich 2017, yielding se- curity guarantees in the standard model which is in contrast to Shi et. al. We show that every mechanism which preserves ǫ, δ -differential pri- p q vacy in effect preserves computational ǫ, δ -differential privacy when it p q is executed through a secure PSA scheme. Furthermore, we introduce a novel perturbation mechanism based on the symmetric Skellam distribu- tion that is suited for preserving differential privacy in the distributed setting, and find that its performances in terms of privacy and accuracy are comparable to those of previous solutions. On the other hand, we leverage its specific properties to construct a computationally efficient prospective post-quantum protocol for differentially private time-series data analysis in the distributed model.
    [Show full text]
  • Probability Distribution Relationships
    STATISTICA, anno LXX, n. 1, 2010 PROBABILITY DISTRIBUTION RELATIONSHIPS Y.H. Abdelkader, Z.A. Al-Marzouq 1. INTRODUCTION In spite of the variety of the probability distributions, many of them are related to each other by different kinds of relationship. Deriving the probability distribu- tion from other probability distributions are useful in different situations, for ex- ample, parameter estimations, simulation, and finding the probability of a certain distribution depends on a table of another distribution. The relationships among the probability distributions could be one of the two classifications: the transfor- mations and limiting distributions. In the transformations, there are three most popular techniques for finding a probability distribution from another one. These three techniques are: 1 - The cumulative distribution function technique 2 - The transformation technique 3 - The moment generating function technique. The main idea of these techniques works as follows: For given functions gin(,XX12 ,...,) X, for ik 1, 2, ..., where the joint dis- tribution of random variables (r.v.’s) XX12, , ..., Xn is given, we define the func- tions YgXXii (12 , , ..., X n ), i 1, 2, ..., k (1) The joint distribution of YY12, , ..., Yn can be determined by one of the suit- able method sated above. In particular, for k 1, we seek the distribution of YgX () (2) For some function g(X ) and a given r.v. X . The equation (1) may be linear or non-linear equation. In the case of linearity, it could be taken the form n YaX ¦ ii (3) i 1 42 Y.H. Abdelkader, Z.A. Al-Marzouq Many distributions, for this linear transformation, give the same distributions for different values for ai such as: normal, gamma, chi-square and Cauchy for continuous distributions and Poisson, binomial, negative binomial for discrete distributions as indicated in the Figures by double rectangles.
    [Show full text]