
130 Knowledge Discovery from Data Streams
Figure 8.6: An illustrative example of a two-layered info-fuzzy network
structure.
is responsible for applying the IN algorithm to the current sliding window of
examples; the Classification Module is responsible for classifying the incoming
examples using the current network; and the Meta Learning Module controls
the operation of the Learning Module. As shown in Figure 8.7, the system
builds a network from T
0
training examples; afterwards, the next V
0
examples
are classified according to the induced network. According to the regenerative
approach of OLIN, when more examples arrive, a completely new network is
built ...