AL-2020/decision_tree.py

185 lines
4.6 KiB
Python
Raw Normal View History

2020-05-19 13:14:19 +02:00
import data
training_data = data.learning_data
header = ['color', 'shape', 'weight', 'size', 'name']
# funkcja która zwraca listę unikalnych wartości z każdej kolumny
def uniqie_vals(rows, col):
return set([row[col] for row in rows])
# zliczamy liczbę wystąpień danego typu w zestawie danych
def class_counts(rows):
counts = {} # label -> count
for row in rows:
name = row[-1]
if name not in counts:
counts[name] = 0
counts[name] += 1
return counts
# funkcja do sprawdzania czy wartość jest wartością numeryczną
def is_numeric(val):
return isinstance(val, int) or isinstance(val, float)
# klasa do zadawania pytań
class Question:
def __init__(self, column, value):
self.column = column
self.value = value
def match(self, example):
val = example[self.column]
if is_numeric(val):
return val >= self.value
else:
return val == self.value
def __repr__(self):
condition = '=='
if is_numeric(self.value):
condition = '>='
return "Is %s %s %s?" % (header[self.column], condition, str(self.value))
def partition(rows, question):
""" podział zbioru informacji
dla każdego rzędu w zbiorze, sprawdź czy zgadza się z pytaniem, jeśli tak
dodaj do 'true' inaczej dodaj do 'false' """
true_rows, false_rows = [], []
for row in rows:
if question.match(row):
true_rows.append(row)
else:
false_rows.append(row)
return true_rows, false_rows
def gini(rows):
""" Gini impurity is a measure of how often a randomly chosen element from
the set would be incorrectly labeled if it was randomly labeled according to
the distribution of labels in the subset. """
counts = class_counts(rows)
impurity = 1
for lbl in counts:
prob_of_lbl = counts[lbl] / float(len(rows))
impurity -= prob_of_lbl ** 2
return impurity
def info_gain(left, right, current_uncertainty):
p = float(len(left)) / (len(left) + len(right))
return current_uncertainty - p * gini(left) - (1 - p) * gini(right)
def find_best_split(rows):
""" znajdź najlepsze możliwe pytanie do zadania, sprawdzając wszystkie
właściwośći oraz licząc dla nich 'info_gain' """
best_gain = 0
best_question = None
current_uncertainty = gini(rows)
n_features = len(rows[0]) - 1
for col in range(n_features):
values = set([row[col] for row in rows])
for val in values:
question = Question(col, val)
true_rows, false_rows = partition(rows, question)
if len(true_rows) == 0 or len(false_rows) == 0:
continue
gain = info_gain(true_rows, false_rows, current_uncertainty)
if gain > best_gain:
best_gain, best_question = gain, question
return best_gain, best_question
class Leaf:
def __init__(self, rows):
self.predicions = class_counts(rows)
class DecisionNode:
def __init__(self, question, true_branch, false_branch):
self.question = question
self.true_branch = true_branch
self.false_branch = false_branch
def build_tree(rows):
gain, question = find_best_split(rows)
if gain == 0:
return Leaf(rows)
true_rows, false_rows = partition(rows, question)
true_branch = build_tree(true_rows)
false_branch = build_tree(false_rows)
return DecisionNode(question, true_branch, false_branch)
def print_tree(node, spacing=""):
if isinstance(node, Leaf):
print(spacing + "Predict", node.predicions)
else:
print(spacing + str(node.question))
print(spacing + '--> True:')
print_tree(node.true_branch, spacing + " ")
print(spacing + '--> False:')
print_tree(node.false_branch, spacing + " ")
def classify(row, node):
if isinstance(node, Leaf):
return node.predicions
if node.question.match(row):
return classify(row, node.true_branch)
else:
return classify(row, node.false_branch)
def print_leaf(counts):
probs = []
for lbl in counts.keys():
probs.append(lbl)
return probs
# my_tree = build_tree(training_data)
#
# print_tree(my_tree)
#
# testing_data = [
# ['gold', 'rectangle', 50, 'medium', 'Name'],
# ['brown', 'rectangle', 55, 'medium', 'Snickers'],
# ['white', 'rectangle', 120, 'big', 'Name']
# ]
#
# test = ['white', 'rectangle', 120, 'big', 'Name']
#
# # for row in testing_data:
# # print(print_leaf(classify(row, my_tree)))
#
# wynik = print_leaf(classify(test, my_tree))[0]
# print(wynik)