随机森林实现代码

该代码使用python实现了，随机森林算法

使用数据库为Adult数据库，可以用来预测居民收入是否超过50K/year。该数据集类变量为年收入是否超过50k，属性变量包含年龄，工种，学历，职业，人种等重要信息，14个属性变量中有7个类别型变量。

数据库操作

引入数据库

import pandas as pd
import numpy as np
import math
import random

df = pd.read_csv('Adult.csv')
df

	age	workclass	fnlwgt	education	education.num	marital.status	occupation	relationship	race	sex	capital.gain	capital.loss	hours.per.week	native.country	income
0	90	?	77053	HS-grad	9	Widowed	?	Not-in-family	White	Female	0	4356	40	United-States	<=50K
1	82	Private	132870	HS-grad	9	Widowed	Exec-managerial	Not-in-family	White	Female	0	4356	18	United-States	<=50K
2	66	?	186061	Some-college	10	Widowed	?	Unmarried	Black	Female	0	4356	40	United-States	<=50K
3	54	Private	140359	7th-8th	4	Divorced	Machine-op-inspct	Unmarried	White	Female	0	3900	40	United-States	<=50K
4	41	Private	264663	Some-college	10	Separated	Prof-specialty	Own-child	White	Female	0	3900	40	United-States	<=50K
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
32556	22	Private	310152	Some-college	10	Never-married	Protective-serv	Not-in-family	White	Male	0	0	40	United-States	<=50K
32557	27	Private	257302	Assoc-acdm	12	Married-civ-spouse	Tech-support	Wife	White	Female	0	0	38	United-States	<=50K
32558	40	Private	154374	HS-grad	9	Married-civ-spouse	Machine-op-inspct	Husband	White	Male	0	0	40	United-States	>50K
32559	58	Private	151910	HS-grad	9	Widowed	Adm-clerical	Unmarried	White	Female	0	0	40	United-States	<=50K
32560	22	Private	201490	HS-grad	9	Never-married	Adm-clerical	Own-child	White	Male	0	0	20	United-States	<=50K

32561 rows × 15 columns

处理缺失值

df1 = df[(df['workclass']!='?')&(df['occupation']!='?')&(df['native.country']!='?')]
df1

	age	workclass	fnlwgt	education	education.num	marital.status	occupation	relationship	race	sex	capital.gain	capital.loss	hours.per.week	native.country	income
1	82	Private	132870	HS-grad	9	Widowed	Exec-managerial	Not-in-family	White	Female	0	4356	18	United-States	<=50K
3	54	Private	140359	7th-8th	4	Divorced	Machine-op-inspct	Unmarried	White	Female	0	3900	40	United-States	<=50K
4	41	Private	264663	Some-college	10	Separated	Prof-specialty	Own-child	White	Female	0	3900	40	United-States	<=50K
5	34	Private	216864	HS-grad	9	Divorced	Other-service	Unmarried	White	Female	0	3770	45	United-States	<=50K
6	38	Private	150601	10th	6	Separated	Adm-clerical	Unmarried	White	Male	0	3770	40	United-States	<=50K
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
32556	22	Private	310152	Some-college	10	Never-married	Protective-serv	Not-in-family	White	Male	0	0	40	United-States	<=50K
32557	27	Private	257302	Assoc-acdm	12	Married-civ-spouse	Tech-support	Wife	White	Female	0	0	38	United-States	<=50K
32558	40	Private	154374	HS-grad	9	Married-civ-spouse	Machine-op-inspct	Husband	White	Male	0	0	40	United-States	>50K
32559	58	Private	151910	HS-grad	9	Widowed	Adm-clerical	Unmarried	White	Female	0	0	40	United-States	<=50K
32560	22	Private	201490	HS-grad	9	Never-married	Adm-clerical	Own-child	White	Male	0	0	20	United-States	<=50K

30162 rows × 15 columns

获取数据集中一列的所有取值，例如sex中的male 和 female

def get_all_category(line):
    line_class=list()
    for i in line:
        if i not in line_class:
            line_class.append(i)
    return line_class

判断两个链表是否相同

def equeal_list(list1,list2):
    if len(list1) != len(list2):
        return False
    for j in range(len(list1)):
        if list1[j] != list2[j]:
            return False
    return True

def equeal_list_not_in_list(list1,list2):
    if len(list2) == 0:
        return True
    for i in range(len(list2)):
        if equeal_list(list1,list2[i]):
            return False
    return True

计算信息增益

计算信息熵

$Ent(x)=-{\sum }_{k=1}^{|y|}{p}_{k}log2(p_k)$

def Ent(data):
    prob=pd.value_counts(data)/len(data)#计算某列不同值，并得出重复值
    return sum(np.log2(prob)*prob*-1)

计算信息增益

$Gain(D,a)=Ent(D)-{\sum }_{v=1}^V\frac{|D^v|}{|D|}Ent(D^v)$

def Gain(data,str1,str2): 
    e1=data.groupby(str1).apply(lambda x:Ent(x[str2]))
    p1=pd.value_counts(data[str1])/len(data[str2])
    e2=sum(e1*p1)
    return Ent(data[str2])-e2

选取最优特征

def get_best_spilt(data,str2):#data为该分支下的数据集，str2为二分类判断特征
    features=list()
    n_features=list()
    for k in data.columns:#所有特征
        n_features.append(k)
    n_features.remove(str2)#除去二分类特征
    k=math.ceil(math.log2(len(n_features)))
    features=random.sample(n_features,k)#取k个特征用来选取最优特征
    l=list()
    for s in features:
        g=Gain(data,s,str2)
        l.append(g)
    return n_features[l.index(max(l))]
    
    

构造决策树类

class Tree():
    def __init__(self,node,depth,is_leaf,spilt,node_df,label,tree_list,all_node):
        self.node=node#当前结点的标签名称----------String
        self.depth=depth#当前结点深度---------int
        self.is_leaf=is_leaf#当前结点是否为叶结点-------int 0 1
        self.spilt=spilt#当前结点的最佳特征--------String
        self.node_df=node_df#当前结点包含的数据集
        self.label=label#当前结点的二分类结果，需要判断是否为叶结点，如果不是叶结点则赋值为2
        self.tree_list=tree_list#当前结点下的所有子树
        self.all_node=all_node#当前已经用了的特征

构造树

决策树的3个中止条件

1、子节点中的样本属于同一类
2、子节点无样本
3、特征已经用完

判断是否为叶结点

def is_leaf_now(data,str2):
    pol=pd.value_counts(data[str2])/len(data[str2])
    if len(data[data.columns[0]])==0:
        return False
    elif pol[0]==1:
        return False
    elif len(data.columns)==1:
        return False
    else:
        return True

建树

def built_tree(data,str2,Tnode,k,all_node):
    now_tree=Tree('',0,0,'',pd.DataFrame(),2,list(),list())#新建一棵树
    now_tree.node=Tnode#当前结点的标签，如male
    now_tree.node_df=data
    now_tree.depth=k+1
#     print('is_leaf_now：',is_leaf_now(data,str2))
    if is_leaf_now(data,str2): 
        now_spilt=get_best_spilt(data,str2)#最佳特征
        now_tree.spilt=now_spilt
        now_tree.all_node=all_node
        #获取最佳特征下的所有取值可能
        all_category=get_all_category(now_tree.node_df[now_tree.spilt])
        for i in all_category:
            now_tree.all_node.append(now_tree.spilt)#为了判断树是否一样
            now_data=now_tree.node_df[now_tree.node_df[now_tree.spilt]==i]#取该属性的所有数据
            del now_data[now_tree.spilt]#删除该特征列的值，不在作为下一步计算
            new_tree=built_tree(now_data,str2,i,now_tree.depth,now_tree.all_node)
            now_tree.tree_list.append(new_tree)
       
    else:
        now_tree.is_leaf=1
        b=str(data[str2].head(1))
        la='<=50K'
        lb='>50K'
        la_result=la in b
        lb_result=lb in b
#         print('la_result：',la_result)
#         print('lb_result：',lb_result)
        if la_result:
            
            now_tree.label=-1
        elif lb_result:
            
            now_tree.label=1
    return now_tree

创建随机森林

def random_forest(train,str2):
    trees=list()
    already=list()
    #先new一颗tree
    Tnode='tree_root'
    depth=0
    all_node=list()
    tree=built_tree(train,str2,Tnode,depth,all_node)
    while len(trees)<=(6*5*4*3*2):
        if equeal_list_not_in_list(tree.all_node,already):#该tree和之前的树没有重复
            already.append(tree.all_node)
            trees.append(tree)
        all_node=list()
        tree=built_tree(train,str2,Tnode,depth,all_node)
    return trees

预测函数

该分支下没有该取值判断

def is_no_var(now_spilt,data,thevar):
    all_catagory=get_all_category(data[now_spilt])
    if thevar not in all_catagory:
        return True
    return False

返回对应值的树

def get_tree_var(thevar,tree):
    for i in range(len(tree.tree_list)):
        if thevar == tree.tree_list[i].node:
            return tree.tree_list[i]

数据进入到Tree得到的结果

def Tree_result(tree,age,workclass,fnlwgt,education,education_num,marital_status,occupation,relationship,race,sex,hours_per_week,native_country):
    now_tree=tree
    label=0
    while now_tree.is_leaf==0:
#         print(now_tree.node_df)
              
        now_spilt=now_tree.spilt
        if now_spilt=='age':
            #如果该分类下没有这个取值返回0
            if is_no_var(now_spilt,now_tree.node_df,age):  
                return 0
            #得到age值对应的树
            now_tree=get_tree_var(age,now_tree)
        elif now_spilt=='workclass':
            if is_no_var(now_spilt,now_tree.node_df,workclass):  
                return 0
            now_tree=get_tree_var(workclass,now_tree)
        elif now_spilt=='fnlwgt':
            if is_no_var(now_spilt,now_tree.node_df,fnlwgt):  
                return 0
            now_tree=get_tree_var(fnlwgt,now_tree)
        elif now_spilt=='education':
            if is_no_var(now_spilt,now_tree.node_df,education):  
                return 0
            now_tree=get_tree_var(education,now_tree)
        elif now_spilt=='eduction.num':
            if is_no_var(now_spilt,now_tree.node_df,education_num):  
                return 0
            now_tree=get_tree_var(education_num,now_tree)
        elif now_spilt=='marital.status':
            if is_no_var(now_spilt,now_tree.node_df,marital_status):  
                return 0
            now_tree=get_tree_var(marital_status,now_tree)
        elif now_spilt=='occupation':
            if is_no_var(now_spilt,now_tree.node_df,occupation):  
                return 0
            now_tree=get_tree_var(occupation,now_tree)
        elif now_spilt=='relationship':
            if is_no_var(now_spilt,now_tree.node_df,relationship):  
                return 0
            now_tree=get_tree_var(relationship,now_tree)
        elif now_spilt=='race':
            if is_no_var(now_spilt,now_tree.node_df,race):  
                return 0
            now_tree=get_tree_var(race,now_tree)
        elif now_spilt=='sex':
            if is_no_var(now_spilt,now_tree.node_df,sex):  
                return 0
            now_tree=get_tree_var(sex,now_tree)
        elif now_spilt=='hours.per.week':
            if is_no_var(now_spilt,now_tree.node_df,hours_per_week):  
                return 0
            now_tree=get_tree_var(hours_per_week,now_tree)
        elif now_spilt=='native.country':
            if is_no_var(now_spilt,now_tree.node_df,native_country):  
                return 0
            now_tree=get_tree_var(native_country,now_tree)
        
#     print(now_tree.label)   
    return now_tree.label

数据进入到random forest 中得到的结果

def random_forest_result(trees,age,workclass,fnlwgt,education,education_num,marital_status,occupation,relationship,race,sex,hours_per_week,native_country):
    k=0
    for i in range(len(trees)):
#         print(Tree_result(trees[i],age,workclass,fnlwgt,education,education_num,marital_status,occupation,relationship,race,sex,hours_per_week,native_country))
        k=k+Tree_result(trees[i],age,workclass,fnlwgt,education,education_num,marital_status,occupation,relationship,race,sex,hours_per_week,native_country)
    
    return k

暂时去掉一些特征，缩小数据集，用来测试程序

del df1['capital.gain']
del df1['capital.loss']
del df1['education.num']
del df1['native.country']
del df1['race']
del df1['hours.per.week']
del df1['sex']
del df1['fnlwgt']
data=df1.head(100)
data

	age	workclass	education	marital.status	occupation	relationship	income
1	82	Private	HS-grad	Widowed	Exec-managerial	Not-in-family	<=50K
3	54	Private	7th-8th	Divorced	Machine-op-inspct	Unmarried	<=50K
4	41	Private	Some-college	Separated	Prof-specialty	Own-child	<=50K
5	34	Private	HS-grad	Divorced	Other-service	Unmarried	<=50K
6	38	Private	10th	Separated	Adm-clerical	Unmarried	<=50K
...	...	...	...	...	...	...	...
108	50	Private	Bachelors	Married-civ-spouse	Exec-managerial	Wife	>50K
109	47	Private	Bachelors	Married-civ-spouse	Exec-managerial	Husband	>50K
110	47	Self-emp-inc	Prof-school	Married-civ-spouse	Prof-specialty	Husband	>50K
111	67	Private	Bachelors	Widowed	Exec-managerial	Not-in-family	>50K
112	67	Self-emp-inc	Bachelors	Married-civ-spouse	Exec-managerial	Husband	>50K

100 rows × 7 columns

训练模型

trees=list()
trees=random_forest(data,'income')
print(len(trees))

721

预测

k=random_forest_result(trees,82,'Self-emp-inc','fnlwgt','Some-college','education_num','Separated','Exec-managerial','Own-child','race','sex','hours_per_week','native_country')
# print('OK!!!')
print(k)

74