Limited Offer Get 25% off — use code BESTW25
No AI No Plagiarism On-Time Delivery Free Revisions
Claim Now
Prices start at $5 per page · plagiarism-free papers · free revisions for 14 days 24/7 support: support@bestwriters.com
Uncategorized

DSC 425 Time Series Analysis and Forecasting

  • Posted
  • Authoradmin
  • Length1,023 words
  • Reading time5 min
Write My Assignment FREE

DSC 425 Time Series Analysis and Forecasting
Fall 2019/2020
Exam
Name:
ID:
Write out the pledge and sign your name.
“I affirm that I have not given or received any unauthorized help on this assignment, and that
this work is my own.”
Signature ________________________
Problem 1 [10 points]
When both seasonal and first differences are applied, it makes no difference which is done first—the
result will be the same. True or False? Provide mathematical proof (If this sounds complicated – and
believe me it is not – you can also show it with an example where you assume first difference (lag 1) and
seasonal difference of some value say 12 (lag 12).
Problem 2 [10 points]
Is it possible to write a stationary AR(p) model as an MA(∞) model? Prove this for the case of AR(1)
model.
Problem 3 [10 points]
Define when a time series process is weakly stationary and explain what the stationarity assumption
implies in terms of modeling and forecasting.
Problem 4 [10 points]
A time series linear regression model was fit to daily data and the following model was obtained. The
demand for electricity is being modeled based on trend, season, temperature and whether it is work day
or not. Assume season1 is for Monday. The data that was used to fit the model started on January 1,
meaning January 1 was the first day.
1. Write the general formula of the model.
2. Write the detailed expression of the model with the proper values.
3. What is the prediction for demand on January 10 2019?
4. What is the prediction for demand on February 10 2019?
Problem 5 [20 points]
The file ”women_murders.csv” contains the number of women murdered each year (per 100,000
standard population) in the United States from 1960 to 2004.
1. Import the data and make the necessary conversion to make it time series. Pay attention that
this is annual data meaning the frequency is 1.
2. Plot the data. What do you see? Any trend, seasonality?
3. Based on the plot is this data stationary? What differences are needed to reach stationarity?
4. By using kpss.test() function within the R “tseries” package identify whether the series
is stationary. Apply several differences until this test shows the data is stationarized. How many
differences did you need to do? Plot the differenced series and comment on it.
5. Now check the ACF and PACF plots of the differenced time series. What do you observe? What
models are appropriate here? Just based on the ACF and PACF plots which order AR, MA, and
ARMA models would you try at first?
6. Fit these models to the data. Do not include any constants. Write down the expressions for each
model (mathematical equation). Are the coefficients significant in each model? What are the AIC
and BIC values? Which model would you choose?
7. Now examine the residuals of each model. Are the models a good fit to the data? Based both on
residual analysis and the AIC metric, which model would you prefer?
8. By using the best model obtained from part 7, create forecasts for the next 6 periods and plot it.
What do you observe?
9. Does auto.arima() give the same model you have chosen? If not, which model do you think is
better?
Problem 6 [40 points]
The file “australia_visitors.csv” contains data monthly Australian short-term overseas visitors between
1990 and 2005.
1. Import the data and make the necessary conversion to make it time series.
I don’t want you mess up here (because everything else will be wrong if you do) so here is the code for
importing and converting:
x = read.csv(‘australia_visitors.csv’)
x = ts(x, start=c(1990,1), frequency=12)
2. Plot the data. What do you see? Any trend, seasonality?
3. Based on the plot is this data stationary? What differences are needed to reach stationarity?
4. Fit a time series linear model using only trend. Plot the original data and the fitted values on the
same plot. Does it look like a good fit?
5. Fit a time series linear model using trend and seasonality. Plot the original data and the fitted
values on the same plot. Does it look like a good fit?
6. Fit a new linear regression model including trend and seasonality using the data from 1990 up
until 2000, including 2000. The following questions will be based on this model.
a. Write the mathematical expression for the linear regression model.
b. Which coefficients are significant?
c. Assuming the smaller the p-value the more significant that variable is, which variable(s)
is the most significant?
d. Now examine the residuals of the model. Is the model a good fit to the data? Based
both on residual analysis what could you say?
e. Forecast for the next 5 years until 2005 April. 2005 April is the end of the time series.
Plot original data and the forecasts. How do the forecasts look like? Visually can you tell
if the model is performing well?
7. By using kpss.test() function within the R “tseries” package identify whether the series
is stationary. Apply several differences until this test shows the data is stationarized. How many
differences did you need to do? Does the data need any seasonal differencing? Plot the
differenced series and comment on it.
8. Now check the ACF and PACF plots of the differenced time series. What do you observe? What
models are appropriate here? Which model is the most appropriate based on the time plot of
the data and the ACF and PACF plots? Write the order of the model.
9. Fit the (SARIMA) model. Perform the residual analysis of the model. Is the model a good fit to
the data?
10. Forecast for the next 5 years until 2005 April (2005 April is the end of the time series) using the
SARIMA model. Plot original data and the forecasts. How do the forecasts look like? Visually can
you tell if the model is performing well?
11. Compare the final model of time series with the SARIMA model visually. Which one is better?
12. Since you have both the actual values and the forecasts for the same periods, you can calculate
the error metrics.
a. Calculate the RMSE and MAPE metrics for the forecasting period for the linear
regression model.
b. Calculate the RMSE and MAPE metrics for the forecasting period for the SARIMA model.
c. Compare these metrics of the two models. Based on these metrics which model is
better?

The post DSC 425 Time Series Analysis and Forecasting appeared first on My Assignment Online.

Plagiarism Free Assignment Help

Expert Help With This Assignment — On Your Terms

✓ Native UK, USA & Australia writers ✓ Deadline from 3 hours ✓ 100% Plagiarism-Free — Turnitin included ✓ Unlimited free revisions ✓ Free to submit — compare quotes

admin

Academic writer and editor at BestWriters.

Get your paper off your plate tonight

Post your instructions for free. Prices start at $5 per page.

Place an order