Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

STA483/583: Take-Home Exam 1

The time-series data named retail is built as part of the TSA library in R. This data set represents the retail sale (in billions) for the United Kingdom for the period January 1986 through March 2007. You will analyze this dataset and forecast the retail sales using the tools in R (introduced in LABs).

Please submit 2 files via Canvas: RDM and PDF/HTML. Late submissions will not be accepted.  Each part below will be graded as: full mark – X pts, partial mark- X/2 pts., no mark – 0 pts.). Partial marks will be given to incomplete answers. Incorrect answers or omissions will earn zero points. Full mark should include complete and correct answers, nice presentation, nice style, and a summary of findings and observations as required below.

The estimated average time for this take-home exam is about 3 hrs.

Import the retail data in R. Remove the first 3 years (3x12=36 observations) from this dataset and continue to work with the remaining observations (i.e., shorter dataset).

1. (6pts) Provide the analysis using basic visualization tools that you learned as part of your previous LABs. A minimum of 3 plots are required for this analysis.  Provide your overall discussion/observations about this TS data using minimum 5 full sentences. 

2. (6pts) Is this data stationary? Why or why not? Discuss using a minimum 5 full sentences.  Make sure to link your discussion to the theoretical part of this TS course (e.g. lecture notes).  

3. (6pts) Can we run this data set using seasonal means model with or without an intercept (the model that we used for the temperature data (Pgs. 32-33 from your book.)? If yes, then run the model and discuss your results. If no, make the appropriate adjustments, run the correct model in R and discus your results. Discussion should include a minimum 5 full sentences.

4. (6pts) Can we run the same data set using the Cosine trend model discussed on Pgs 34-35 of your book? Justify without producing any analysis in R.  

5. (7 pts) Partition the data into training set and test set. Develop a model that should be used for modeling this type of data. Your analysis should include the exponential smoothing model and a variety of Holt-Winters models (both additive and multiplicative).  A minimum of 3 visual plots are required as part of this analysis. Propose the best model. Discuss your observations and the results using a minimum 5 full sentences.  

6. (7pts) Forecast the retail sales for 4 periods in the future. As part of this task develop a minimum of 2 visual plots. Report your forecast and discuss your results using a minimum 5 full sentences.

7. (6pts) The RDM Report is free of any clutter (e.g. warning messages, unnecessary output appears in the report from loading R packages or running accidentally the entire data frame object) and it is nicely formatted. All visual plots are labeled properly and the report is free of spelling and grammar errors. No running text appears in the plotting area.

8. (6pts) Reflect on the in-class and take-home portion of the exam (a min 5 full sentences are required).