| Select a data set
{x1, x2, ... , xn} |
| and possibly preprocess it: derive the related data sets, for example |
| {x2 - x1, x3 - x2, ... }
(difference delay = 1, offset = 1), |
| one-step differences | |
| {x3 - x1, x4 - x2, ... }
(difference delay = 2, offset = 1), |
| two-step differences with overlap | |
| {x3 - x1, x5 - x3, ... }
(difference delay = 2, offset = 2) |
| two-step differences without overlap | |
|
| Run these data sets through a
driven IFS program, taking care to select appropriate bin boundaries. |
| Interpret the
results, looking for forbidden combinations and more commonly occurring
combinations. |
| Translate your observations back into the context of the original data set.
|