Now we have these descriptions of various events, what can we do with them? Since each PEP is described in terms of a set of parameters, we can now do clustering in the parameter space. For example, if we look at the ``delta'' PEP, we can do clustering in two dimensions to find groups of event primitives which have similar parameter values, and thus have similar characteristics. For example, if we take the data from table 5.2 and plot them (as shown in figure 5.5) then we can see that we might be able to form several possible clusters in the parameter space to represent different events. Each of these clusters represents a particular type of event, which we will term a synthetic event.
Figure 5.5: Parameter space and clustering for Blues and Reds
domain.
Thus what we will get from the clustering is some set of clusters representing different types of events, even though they are of the same type of primitive.