This is best understood with an example. Suppose we have the following sets of equally spaced data:
1. Monthly degree days recorded for a city in a given year (12 points - data set 1).
2. The gas used by the city during each month. in year 1 (12 points data set 2).
3. The prediction of the monthly degree days for year 2 (12 points data set 3).
We can prdict the gas used per month in the upcoming year 2, by inputting data sets one and two as curves B and C and doing a deconvolution providing the response function as curve A (A=C/B). Now, enter the third set of data as B, and do a convolution against the response function and the output, curve C (C=A*B), is the predicted values of the gas usage in the upcoming year.
Well that was the example in the book.
|