Scribd
Upload
55 views

MODULE 3: Stochastic Processes and Time Series

This module provides an extended coverage of stochastic processes including Markov chains and various forms of Poisson processes. It also covers time series modeling including ARIMA modeling. Key topics covered include random walks, branching processes, Markov chains, Poisson processes, birth and death processes, queues like the M/M/1 queue, and time series concepts like stationarity, autocorrelation, ARIMA processes, and forecasting. Applications in substantive areas will be represented in exam questions.

Uploaded by

S.L.L.C
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
55 views

MODULE 3: Stochastic Processes and Time Series

This module provides an extended coverage of stochastic processes including Markov chains and various forms of Poisson processes. It also covers time series modeling including ARIMA modeling. Key topics covered include random walks, branching processes, Markov chains, Poisson processes, birth and death processes, queues like the M/M/1 queue, and time series concepts like stationarity, autocorrelation, ARIMA processes, and forecasting. Applications in substantive areas will be represented in exam questions.

Uploaded by

S.L.L.C
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

MODULE 3: Stochastic processes and time series

This module provides an extended coverage of stochastic processes, including Markov chains and various forms of
Poisson processes, and of time series, including ARIMA modelling.
The syllabus concentrates on underlying theory, but applications in various substantive areas are also important and
will be represented in examination questions.

Stochastic processes
General stochastic process models.
Random walks.
Reflecting and absorbing barriers.
Mean recurrence time, mean time to absorption.
Difference equations.
Branching processes.

Use of generating functions.


Recurrence relations for size of n th generation;
probability of extinction.

Markov chain models for discrete-state processes. Not restricted to finitely many states.
Transition matrices: 1-step and n-step.
Classification of states.
Equilibrium distributions for time-homogeneous
chains.
Poisson processes.
Differential-difference equations.
Birth and death processes.
Queues.
The M/M/1 queue. Differential-difference equations. Conditions for equilibrium. Equilibrium Equilibrium behaviour for queues with transition
distributions of queue size and waiting time for rates dependent on queue size.
first-come-first-served queues. Extensions to
M/M/k and M/M/ queues.
The M/G/1 queue, imbedded Markov chain
analysis.
The Pollaczek-Khintchine formula. Equilibrium treatment only.
Mean queue length and waiting time.

Time series
Time series models; trend and seasonality.

In examination questions, the word "queue" will


refer to all units in a system, i.e. those being
served as well as those still waiting to be served.
Additive and multiplicative models.

Stationarity.
Autocovariance, autocorrelation and partial
autocorrelation functions. Correlograms.
Autoregressive (AR) processes.
Yule-Walker equations.
Moving average (MA) processes.
Invertibility conditions.
ARMA processes.
ARIMA processes and Box-Jenkins methods.
Identification, estimation, checking, forecasting.

Box-Pierce and Ljung-Box statistics.

Forecasting and minimising expected prediction Exponential smoothing, Holt-Winters.


variance.
Introduction to frequency domain analysis. Candidates will be expected to have some
Spectral density function. Periodograms.
familiarity with the fast Fourier transform.
28

You might also like