ISYE 6402 Exam 3: Time Series Analysis Verified
and Latest Questions and Answers - Georgia Tech
1. In an AR(p) process, what is the characteristic behavior of the Partial
Autocorrelation Function (PACF)?
A. It cuts off after lag p.
B. It tails off exponentially or follows a sine wave pattern.
C. It is zero for all lags.
D. It cuts off after lag 1 regardless of p.
Answer: A
Explanation: For an AR(p) process, the PACF cuts off after lag p, while the ACF tails off.
This property is used to identify the order of the autoregressive component.
2. Which of the following is a primary requirement for a time series to be
considered ‘weakly stationary’?
A. The mean is zero at all times.
B. The series must have no seasonal components.
C. The series must follow a normal distribution.
D. The variance is constant and the autocovariance depends only on the lag.
Answer: D
Explanation: Weak stationarity requires a constant mean and a variance that does not
change over time, and the autocovariance between two time points depends only on the
distance (lag) between them.
,3. When using the Box-Jenkins methodology, what does the ‘I’ in ARIMA(p, d, q)
represent?
A. Integrated
B. Independent
C. Invertible
D. Interrupted
Answer: A
Explanation: The ‘I’ stands for Integrated, referring to the differencing of the raw
observations to allow for the time series to become stationary.
4. In a GARCH(1,1) model, what does the model primarily aim to capture?
A. The linear trend of the series.
B. The seasonal periodicity.
C. The relationship between two independent variables.
D. Time-varying conditional volatility.
Answer: D
Explanation: GARCH (Generalized Autoregressive Conditional Heteroskedasticity) models
are designed to model and forecast conditional variance, capturing ‘volatility clustering’.
5. Which criterion penalizes the number of parameters more heavily, often
leading to simpler models?
A. AIC (Akaike Information Criterion)
B. BIC (Bayesian Information Criterion)
C. R-squared
D. Mean Absolute Error
Answer: B
Explanation: BIC applies a larger penalty for the number of parameters (specifically
involving the log of the sample size) compared to AIC, which tends to favor more
parsimonious models.
, 6. What is the purpose of the Ljung-Box test in time series analysis?
A. To test if the series is stationary.
B. To determine the order of differencing.
C. To check if the residuals are white noise.
D. To find the optimal seasonal period.
Answer: C
Explanation: The Ljung-Box test is a portmanteau test used to examine whether any of a
group of autocorrelations of a time series (usually residuals) are different from zero.
7. In Spectral Analysis, the periodogram is an estimate of which of the
following?
A. The moving average coefficients.
B. The cumulative distribution function.
C. The spectral density function.
D. The partial autocorrelation.
Answer: C
Explanation: The periodogram is used to identify the dominant periods (or frequencies) of
a time series by estimating the spectral density.
8. A SARIMA(0,0,1)x(0,0,1)_12 model is equivalent to which process?
A. An MA process with lags at 1, 12, and 13.
B. A pure MA(1) process.
C. An AR(13) process.
D. A random walk with drift.
Answer: A
Explanation: Expanding (1 + thetaB)(1 + ThetaB^12) results in terms at lags 1, 12, and 13.
and Latest Questions and Answers - Georgia Tech
1. In an AR(p) process, what is the characteristic behavior of the Partial
Autocorrelation Function (PACF)?
A. It cuts off after lag p.
B. It tails off exponentially or follows a sine wave pattern.
C. It is zero for all lags.
D. It cuts off after lag 1 regardless of p.
Answer: A
Explanation: For an AR(p) process, the PACF cuts off after lag p, while the ACF tails off.
This property is used to identify the order of the autoregressive component.
2. Which of the following is a primary requirement for a time series to be
considered ‘weakly stationary’?
A. The mean is zero at all times.
B. The series must have no seasonal components.
C. The series must follow a normal distribution.
D. The variance is constant and the autocovariance depends only on the lag.
Answer: D
Explanation: Weak stationarity requires a constant mean and a variance that does not
change over time, and the autocovariance between two time points depends only on the
distance (lag) between them.
,3. When using the Box-Jenkins methodology, what does the ‘I’ in ARIMA(p, d, q)
represent?
A. Integrated
B. Independent
C. Invertible
D. Interrupted
Answer: A
Explanation: The ‘I’ stands for Integrated, referring to the differencing of the raw
observations to allow for the time series to become stationary.
4. In a GARCH(1,1) model, what does the model primarily aim to capture?
A. The linear trend of the series.
B. The seasonal periodicity.
C. The relationship between two independent variables.
D. Time-varying conditional volatility.
Answer: D
Explanation: GARCH (Generalized Autoregressive Conditional Heteroskedasticity) models
are designed to model and forecast conditional variance, capturing ‘volatility clustering’.
5. Which criterion penalizes the number of parameters more heavily, often
leading to simpler models?
A. AIC (Akaike Information Criterion)
B. BIC (Bayesian Information Criterion)
C. R-squared
D. Mean Absolute Error
Answer: B
Explanation: BIC applies a larger penalty for the number of parameters (specifically
involving the log of the sample size) compared to AIC, which tends to favor more
parsimonious models.
, 6. What is the purpose of the Ljung-Box test in time series analysis?
A. To test if the series is stationary.
B. To determine the order of differencing.
C. To check if the residuals are white noise.
D. To find the optimal seasonal period.
Answer: C
Explanation: The Ljung-Box test is a portmanteau test used to examine whether any of a
group of autocorrelations of a time series (usually residuals) are different from zero.
7. In Spectral Analysis, the periodogram is an estimate of which of the
following?
A. The moving average coefficients.
B. The cumulative distribution function.
C. The spectral density function.
D. The partial autocorrelation.
Answer: C
Explanation: The periodogram is used to identify the dominant periods (or frequencies) of
a time series by estimating the spectral density.
8. A SARIMA(0,0,1)x(0,0,1)_12 model is equivalent to which process?
A. An MA process with lags at 1, 12, and 13.
B. A pure MA(1) process.
C. An AR(13) process.
D. A random walk with drift.
Answer: A
Explanation: Expanding (1 + thetaB)(1 + ThetaB^12) results in terms at lags 1, 12, and 13.
