Information Technology Reference
In-Depth Information
to the frame model, and can be expanded if needed. The controller for rule
production combine the atom sets according to the requirements of experts,
then it fills the blackboard with these combinations and offers two kinds of
functions(blocking and enumerating) to promote the inference efficiency,
finally produces the particular rules. This manner offers the ability of rule
production gradually, lighten the inference load of expertise reasoning, and
avoid the omission of knowledge while keeping focus on the internal structure
of knowledge. The visual interface that the system used to express such
structure enhances the knowledge expression method and offers effective
inference ability.
Model Three: Fuzzy reasoning. We use fuzzy methods to choose the
rules to be used in the revision of prediction results and thus more accurate
results can be generated. In the rule set of prediction and revision system, the
positions, sizes, outputs of central fishing grounds(X axle, Y axle) are defined
fuzzily, and new predication results can be reached by revising the results
generated by CBR reasoning. After the whole predication process, we can
revise the conclusion sets through a learning process on the sequence of
conclusions ordered by some pre-defined credibility degrees. In this manner,
we can approach constantly and gradually to a believed solution.
Fig. 5.7. System architecture
