biggestt / id3-decision-tree Goto Github PK

Attempt to implement the ID3 decision tree algorithm in Octave. Takes a .csv file as input and prints tree to console. Currently no penalty for multi-variate attributes so I suggest you only use binary valued attributes.

Example of using training data from example.csv with the first column as target attribute and the rest as input attributes:

octave-cli
octave:1> id3('example.csv', 1, [2:4])
employed value: 0
   rich value: 0
      not happy
   rich value: 1
      married value: 0
         happy most likely (50%)
      married value: 1
         happy
employed value: 1
   not happy