python类概念是什么_python 类定义

classNode(object):def__init__(self,X,y,split=Split.null()):self.X=Xself.y=yself.split=splitself.left=Noneself.right=Noneself.outcome=np.mean(self.y)ifself.y.any()elseNone...

class Node(object):

def __init__(self, X, y, split=Split.null()):

self.X = X

self.y = y

self.split = split

self.left = None

self.right = None

self.outcome = np.mean(self.y) if self.y.any() else None

def traverse(self):

result = []

nodes = [self] #这是什么意思？？

while nodes:

current_node = nodes.pop()

result.append(current_node)

if current_node.left:

nodes.insert(0, current_node.left)

if current_node.right:

nodes.insert(0, current_node.right)

return result

def predict(self, X):

result = np.zeros(X.shape[0])

for i, x in enumerate(X):

nodes = [self]

while nodes:

current_node = nodes.pop()

if current_node.is_leaf:

result[i] = current_node.outcome

break

if x[current_node.split.split_attribute] < current_node.split\

.split_value:

nodes.insert(0, current_node.left)

else:

nodes.insert(0, current_node.right)

return result

def fit(self, X=None, y=None, max_tree_size=10):

self.X = X or self.X

self.y = y or self.y

tree = [self]

n_nodes = 1

while len(tree):

if n_nodes >= max_tree_size:

break

node = tree.pop()

node.grow()

if node.left:

n_nodes += 1

tree.insert(0, node.left)

if node.right:

n_nodes += 1

tree.insert(0, node.right)

return self

def grow(self, verbose=0):

self.outcome = np.mean(self.y) if not self.outcome else self.outcome

self.split = split(self.X, self.y)

if self.split == Split.null():

return

indexes = self.split.indexes(self.X)

self.left = Node(self.X[indexes], self.y[indexes])

self.right = Node(self.X[~indexes], self.y[~indexes])

@property

def is_leaf(self):

return not (self.right or self.left)

def __str__(self):

return '' % (

#self.X.shape if self.X != None else None,

self.X.shape if self.X.any() else None,

self.split.split_attribute,

self.split.split_value,

self.outcome)

@property

def pstr(self):

return '\n'.join([str(n) for n in self.traverse()])

nodes = [self] #这是什么意思？？

展开