In this paper, the Fuzzy Meta-Learning methodology for
distributed data mining was introduced and preliminary
results from its application to real-world data were presented.
As the experimental results exhibited, it leads to
models of collective knowledge with increased accuracy.
The use of the proposed schemes that take into consideration
the uncertainty of the base models has repeatedly led
to meta-models with increased predictive performance.
The use of a fuzzy inductive algorithm for meta-learning
has also shown promising signs in terms of accuracy. Although
no general claims can at this point be made about
the efficiency of this approach due to limited experiments,
the results are encouraging for further research.