random-forest.yml

metadata:
  title: Predictive analytics
  documentation: "https://docassemble.org/docs/ml.html#RandomForestMachineLearner"
  example start: 1
  example end: 2
---
mandatory: True
code: |
  rf = RandomForestMachineLearner('fruit', use_initial_file=True)
  if rf.is_empty():
    rf.add_to_training_set({'a': 1, 'b': 2.0, 'c': 3.0, 'd': 'A'}, 'apple')
    rf.add_to_training_set({'a': 0.9, 'b': 2, 'c': 3.0, 'd': 'A'}, 'apple')
    rf.add_to_training_set({'a': 0.8, 'b': 2.0, 'c': 3.0, 'd': 'A'}, 'apple')
    rf.add_to_training_set({'a': 1.0, 'b': 2.1, 'c': 3, 'd': 'B'}, 'apple')
    rf.add_to_training_set({'a': 1.2, 'b': 2.0, 'c': 3.1, 'd': 'A'}, 'apple')
    rf.add_to_training_set({'a': 1.0, 'b': 7.0, 'c': 3.0, 'd': 'B'}, 'orange')
    rf.add_to_training_set({'a': 0.9, 'b': 6.9, 'c': 3.0, 'd': 'A'}, 'orange')
    rf.add_to_training_set({'a': 0.8, 'b': 6.7, 'c': 2.9, 'd': 'B'}, 'orange')
    rf.add_to_training_set({'a': 1.0, 'b': 7.1, 'c': 2.95, 'd': 'B'}, 'orange')
    rf.add_to_training_set({'a': 1.05, 'b': 6.8, 'c': 3.1, 'd': 'B'}, 'orange')
---
code: |
  sample_data = {'a': 1, 'b': 3.8, 'c': 3.1, 'd': 'B'}
  predictions = rf.predict(sample_data, probabilities=True)
---
mandatory: True
question: |
  The prediction
subquestion: |
  With
  ${ '{0:.1f}%'.format(predictions[0][1]*100) }
  certainty, I think it is an
  ${ predictions[0][0] }.