Quantitative Forecasting

Quantitative forecasting projects history into the future.  In other words, it tries to understand what happened in the past and use that to predict the future.  Accordingly, quantitative forecasting involves two steps:

1. Modeling the past
2. Using the model to predict the future

Modeling the past involves plotting the data on a scatter diagram to fit one of three models:

1. Constant, x_t = a + e_t , e_t ~NID(0, s²)
2. Linear, x_t = a + bt + e_t , e_t ~NID(0, s²)
3. Seasonal, x_t = (a + bt) c_t + e_t , or x_t = a + bt + c_t + e_t, e_t ~NID(0, s²)

Once the model is determined (Graphically), forecasting may proceed as follows:

For the constant model, use moving averages or exponential smoothing

Summary

Example

For the linear model, use least squares, double exponential smoothing, double moving averages.

Summary

Example