Statistics 5550: Homework #1
Due Friday, January 30, 2015
All problems are from the Shumway & Stoffer text (E-Z version) unless otherwise specified.
Your write up (hard copy) that integrates your solutions with any figures and (formatted)
computer output

Statistics 5550
Introductory Time Series Analysis
Spring 2015 Syllabus
Instructor:
Email:
Office:
Website:
Dr. Christopher Hans
hans@stat.osu.edu
327 Cockins Hall
Carmen
When:
Where:
Office Hours:
MWF 10:20-11:15
CL 135
Mondays 12:30-1:30
Thursdays 12:30-

_,_-
Statistics 5550: Midterm 2
March 30, 2016
_._._._._-
Instructions: There are four questions on this exam, each with several parts. Record your
solutions in the spaces provided. You must show enough work to make your method of
solution clear. If you u

Statistics 5550: lVIidterH'i 1
February 24, 2016
5: There are three questions on this exaIn, each with several parts. Record
15 in the spaces provided. You 1111181; Show enough \vork to Blake your Inethod of
r. If you use a, calculator to Inake a calcul

Reading: 1.5
Stationary Time Series
We have seen features of time series models and data that can change over time:
mean of the series (trend and/or seasonality)
variance (J & J quarterly earnings)
covariance/correlation (random walk)
The fewer aspects

Reading: 1.3
Agenda
We will discuss several basic examples of time series that will be used as building blocks for the more
complex models we will explore later in the semester.
White Noise
white noise
iid noise
Gaussian white noise
Moving Averages

Reading: 1.6
Stationary Time Series
So far we have started with a time series model and derived properties of the model
mean function
ACVF
stationarity
etc.
When analyzing time series data, we start with a data set and estimate aspects of a model that des

Reading: 2.1
Linear Regression Model
A linear regression model describes the relationship between a set of "predictor" variables \(z_cfw_t1,
z_cfw_t2, \ldots, z_cfw_tq\) and a RV \(x_t\) as: \[ x_t = \beta_1 z_cfw_t1 + \beta_2 z_cfw_t2 + \cdots + \beta_q