机器学习手撕代码（2）决策树及随机森林

• 本篇分享一下决策树及随机森林的代码，DTandRF.py为决策树以及随机森林两个模型的代码。utils.py中为可视化结果的工具。
• dataset见本系列第0篇。

DTandRF.py

import numpy as np
from datasets.dataset import DataSet
from sklearn.model_selection import train_test_split
from utils import Visualization,Metrics

def Gini(targets):
    N = len(targets)+1e-5
    count = np.bincount(targets)
    return (1-((np.sum(count**2))/(N**2)))

class Node:
    def __init__(self,max_sections=10):
        self.type = None
        self.threshold = None
        self.feat = None
        self.left = None
        self.right = None
        self.label = None
        self.max_sections = max_sections # 寻找最优阈值的时候最多尝试max_sections次
    def fit(self,data,targets,stop_n):
        count = np.bincount(targets)
        condition = len(data)<=stop_n or np.max(count)/np.sum(count)>0.99 # 是否终止分裂
        if condition: # 叶节点，选择流入当前节点的数据中类别最多的类作为本叶子节点的标签
            self.type = 'leaf'
            self.label = np.argmax(count)
        else:
            self.type = 'root'
            best_feat_i = 0
            best_hold = 0
            best_l_i = None
            best_r_i = None
            G = None
            for i in range(len(data[0])):
                feat_i = data[:,i]
                # 如果流入当前节点的数据小于max_sections，则阈值在这些数据中寻找，否则等距取# 如果流入当前节点的数据小于max_sections个数并寻找最优阈值
                if len(targets)<=self.max_sections:
                    holds = feat_i
                else:
                    holds = np.linspace(np.min(feat_i),np.max(feat_i),self.max_sections)
                for hold in holds:
                    l_i = (feat_i<=hold)
                    r_i = (feat_i>hold)
                    new_g = (len(l_i)*Gini(targets[l_i])+len(r_i)*Gini(targets[r_i]))/len(targets)
                    if G is None or new_g<G:
                        G = new_g
                        best_feat_i = i
                        best_hold = hold
                        best_l_i = l_i
                        best_r_i = r_i
            self.feat = best_feat_i
            self.threshold = best_hold
            self.left = Node(self.max_sections) # 递归建立左子树
            self.left.fit(data[best_l_i],targets[best_l_i],stop_n)
            self.</