Reference no: EM133082624

Forecasting techniques: workshop on advanced models

Learning Outcome 1: Describe an example of a financial problem

Learning Outcome 2: Overview of Data

Learning Outcome 3: Analyse the data

Learning Outcome 4: Evaluate results collaboratively in class

Instruction: Form small groups

Instructions:

• Access the dataset in file ‘M2CALL.xls'

• The spreadsheets contain the following observations
- Name of the series
- Length of seasonality (12 for Monthly, 4 for Quarterly)
- Starting and Ending dates
- Number of observations
- N Observations

• Open the spreadsheet labelled ‘M2C88All'
- Choose a company that has a monthly series with 82 data points (make sure you don't pick an index, check with your facilitator)
- Each group should pick a different company

• Create a separate spreadsheet with the data for the company of your choice (You chose how you represent the data to do a time series analysis).

• Add relevant data from the spreadsheet M2CAll, so that you can determine how the forecasting method performed.

• Load the spreadsheet you created into Exploratory or Orange

1. Plot the original company data (months 1 to 82)

2. Forecast the value for the next 14 months and plot it on the same graph if possible.

Questions

Q1. What tool and method did you use to forecast? Why?

Q2. Can you plot the actual values (for months 83-96) on the same plot?

Q3. Are you able to get confidence intervals? If so are the real values within the confidence intervals of the forecast?

• Discuss how the forecast compares to reality.
• An ARIMA model was created to forecast the forecast the next time periods.