In this chapter, we reviewed the following three methodologies: Markov models, ARIMA, and MCMC as part of Proc MI. We also reviewed various terms such as stationarity, trend, autocorrelation, residual plots, and so on.

We have learned that Markov models can be used both for forecasting and imputation. We compared our results to ARIMA process. Using an alternate scenario for the transition matrix, we have shown how easy it is to come up with forecasts based on different assumptions about transition states.

In our business problem, we have tried to help the finance team come up with robust projections about the number of customers they can expect to have over the next two and half years across various customer accounts. The finance team can use this information to estimate the revenue generated. Along with the operating costs and other inputs available to the finance team, they can compute the profitability of the customer accounts. In the next chapter, we will cover econometric modeling for inflation forecasting.