ARIMA model approach

To explore the basic concepts of ARIMA, please refer to Chapter 2, Forecasting Stock Prices and Portfolio Decisions Using Time Series. In this section, we will try and find the best suited ARIMA model to our three series: classic, premium, and platinum account types. We will be using historic time series data (Figure 4.3) to generate forecasts for five periods.

This is the data setup and ARIMA identification for the classic account:

Data Past; 
input Cl Pr Pl; 
datalines; 
30.68 5.73 1.51 
30.65 5.74 1.53 
30.83 5.43 1.66 
30.9 5.3 1.72 
31.1 4.7 2.12 
31.05 4.73 2.14 
31.01 4.81 2.1 
30.7 5.01 2.21 
30.3 5.3 2.32 
29.3 6.4 2.22 
29.3 6.5 2.12 
28.8 7.3 1.82 
28.8 8.1 1.02 
28.7 8.3 0.92 
28.6 8.34 0.98 
28.4 8.37 1.15 
27.6 9.01 1.31 
26.5 9.5 1.92 
26 9.8 2.12 
25.3 10.3 2.32 
; 
Run; 
 
Ods Graphics On; 
PROC ARIMA Data=Past; 
identify var=Cl scan esacf...