4.5 Model clusters

Choose this option to create a model for each of your clusters. Such models predict whether an instance is part of its respective cluster. This is very useful to see which fields and values are most relevant in deciding membership in a given cluster. For example, if your cluster has two relevant fields that differentiate its instances from the rest of the dataset, you will likely get a decision tree model with a rule set that only involves those two fields in deciding membership. (See section 5.2 .)

By default, BigML does not model clusters.