3. Machine learning techniques and our analysis assumptions
There are a large number of ML techniques each with
diVerent strengths and weaknesses. Choosing which is the
most appropriate technique often requires an understanding
of both the problem domain and the diVerent learning
methods. A good introduction to many machine learning
techniques can be found in [10]. The machine learners used
in this analysis were:
MC4 Decision trees. Decision trees provide a visual representation
of relationships which appear to eVect the
situation under investigation. Pruning is generally
used to reduce the size of the tree. The conWdence
method of pruning was used.