[−][src]Function oner_induction::discover

pub fn discover<A, C>(
    attributes: &ArrayView<A, Ix2>, 
    classes: &ArrayView<C, Ix1>
) -> Option<(usize, Rule<A, C>)> where
    A: Eq + Hash + Clone + Debug,
    C: Eq + Hash + Clone + Debug, 

Find the one rule that fits a set of example data points.

Arguments

  • attributes - rows containing attribute values as columns.
  • classes - the true classification for each row.

Result

The result is a tuple containing the column index the best rule applies to, and the Rule itself.

The Rule consists of:

  • the attribute index (from zero) that the rule works for; and
  • the Cases for that attribute.

A Case is a value for the attribute and the corresponding predicted class.