Tp. Hong et al., TRADE-OFF BETWEEN REQUIREMENT OF LEARNING AND COMPUTATIONAL COST, IEICE transactions on information and systems, E81D(6), 1998, pp. 565-571
Machine learning in real-world situations sometimes starts from an ini
tial collection of training instances; learning then proceeds off and
on as new training instances come intermittently. The idea of two-phas
e learning has then been proposed here for effectively solving the lea
rning problems in which training instances came in this two-stage way.
Four two-phase learning algorithms based on the learning method PRISM
have also been proposed for inducing rules from training instances. T
hese alternatives form a spectrum, showing achievement of the requirem
ent of PRISM (keeping down the number of irrelevant attributes) heavil
y dependent on the spent computational cost. The suitable alternative,
as a trade-off between computational costs and achievement to the req
uirements, can then be chosen according to the request of the applicat
ion domains.