def test_filter_rows():
data = filter_rows(weather_data, lambda r: value(weather_data, r, 'play'))
assert relation(data) == 'weather.symbolic'
assert attributes(data) == [
'outlook', 'temperature', 'humidity', 'windy', 'play'
]
assert rows(data)[0] == ['overcast', 'hot', 'high', False, True]
def get_best_attribute(data, class_attr):
attrs = list(filter(lambda a: a != class_attr, attributes(data)))
print(f' ---\nDetermining best attribute from: {", ".join(attrs)}')
lowest_attr = attrs[0]
lowest_info = 1
for attr in attrs:
print_info_calc(data, class_attr, attr)
info = attr_info(data, class_attr, attr)
if info < lowest_info:
lowest_attr = attr
lowest_info = info
print(f'Lowest info attribute is: {lowest_attr}')
return (lowest_attr, lowest_info)
def test_attributes():
assert attributes(weather_data) == [
'outlook', 'temperature', 'humidity', 'windy', 'play'
]
def test_remove_attribute():
data = remove_attribute(weather_data, 'outlook')
assert relation(data) == 'weather.symbolic'
assert attributes(data) == ['temperature', 'humidity', 'windy', 'play']
assert rows(data)[0] == ['hot', 'high', False, False]
assert rows(data)[7] == ['mild', 'high', False, False]
from P458.data import (
attributes,
read_arff,
)
from P458.id3 import (
id3, )
from P458.tree import (
str_tree,
decide,
)
data = read_arff('./data/contact-lenses.arff')
decision_tree = id3(data, 'contact-lenses')
row = ['sunny', 'hot', 'high', False, False]
result = decide(decision_tree, row, attributes(data))
print(' --- ')
print(str_tree(decision_tree))
print(f'For row: {row} id3 predicts: {result}')