python分類分析--隨機森林原理及案例

隨機森林

1、什麼是集成學習方法

集成學習通過建立幾個板型組合的來解決單一預測問題,它的工作原理是生成多個分類器/模型,各獨立地學習和作出預測。這些預測最後結合成組合預測,因此優於任何一個單分類的做出預測。決策樹過度擬合可以用剪枝或者集成學習方法的隨機森林實現。

2、什麼是隨機森林

在機器學習中,隨機森林是一個包含多個決策樹的分類器,並且其輸出的類別是由多個決策樹輸出的類別的衆數而定。例如,如果你訓練了5個樹,其中有4個樹的結果是True,1個樹的結果是False,那麼最終投票結果就是True。
隨機:
森林:包含多個決策樹的分類器

3、隨機森林的原理過程

隨機:特值隨機,訓練集隨機
隨機森林算法根據下列算法而建造每棵樹:
·用N來表示訓練用例(樣本)的個數,M表示特徵數目。
    。1、一次隨機選出一個樣本,重紅N次。《隨機有放回的抽取,有可能出現重複的樣本)
    。2、隨機去選出m個特徵,m << M,建立決策制,每棵樹有m個特徵。
·採取bootstrap抽樣 《隨機有放回的抽樣》   

4、爲什麼採取bootstrap抽樣

爲什麼要隨機推樣訓練?

  • 如果不進行隨機抽樣,每棵樹的訓練集都一樣,那麼最終訓練出的樹分類結果也一樣

爲什麼要有放回地抽樣?

  • 如果不是有放回的抽樣,那麼每棵樹的訓練樣本都是不同的,都是沒有交集的,也就是說每棵樹訓練出來都是有很大的差異的;而隨機森林最後分類取決於多棵樹(弱分類器》的投票表決。

5、python實現隨機森林的接口

· class sklearn.ensemble.RandomForestClassifier(n_estimators=10,criterion='gini',max_depth=None,bootstrap=True,random_state=None,min_samples_split=2)

隨機森林分類器參數解釋:

。n_estimators:integer,optional(default=10)森林裏的樹木數量100,150,300,...
。criteria:string,可選(default="gini")分割特徵的測量方法"entropy"、"gini"
。max_depth:integer或None,可選(默認=無)樹的最大深度5,8,15,25,30
。max_features="auto”,每個決策樹的量大特徵數量
    ·If"auto",then max_features=sqrt(n_features).
    ·If"sqrt",then max_features=sqrt(n features)(same as"auto").
    ·If"log2",then max_features=log2(n_features).
    ·If None,then max_features = n_features.
。bootstrap:boolean,optional(default=True)是否在構建樹時使用放回抽樣\
。min_samples_split:節點劃分最少樣本數
。min_samples_leaf:葉子節點的最小樣本數

·其中超參數有:n_estimator,max_depth,min_samples_split,min_samples_leaf

6、應用總結

在當前所有分類算法中,具有極好的準確率

能夠有效地運行在大數據集上,處理具有高維特徵的輸入樣本,而且不需要降維

能夠評估各個特徵在分類問題上的重要性

7、案例:隨機森林對泰坦尼克號乘客的生存進行預測

import pandas as pd

'''1 獲取數據'''
path = "http://biostat.mc.vanderbilt.edu/wiki/pub/Main/DataSets/titanic.txt"
titanic = pd.read_csv(path)
row.names pclass survived name age embarked home.dest room ticket boat sex
0 1 1st 1 Allen, Miss Elisabeth Walton 29.0000 Southampton St Louis, MO B-5 24160 L221 2 female
1 2 1st 0 Allison, Miss Helen Loraine 2.0000 Southampton Montreal, PQ / Chesterville, ON C26 NaN NaN female
2 3 1st 0 Allison, Mr Hudson Joshua Creighton 30.0000 Southampton Montreal, PQ / Chesterville, ON C26 NaN (135) male
3 4 1st 0 Allison, Mrs Hudson J.C. (Bessie Waldo Daniels) 25.0000 Southampton Montreal, PQ / Chesterville, ON C26 NaN NaN female
4 5 1st 1 Allison, Master Hudson Trevor 0.9167 Southampton Montreal, PQ / Chesterville, ON C22 NaN 11 male
display(titanic.head(3))
# 解釋數據
#字段: row.names  	pclass  	survived	name	age	embarked	home.dest	room	ticket	boat	sex
#解釋:人員編號    人員等級劃分  是否倖存    名字  年齡  上傳地點  家庭地址  所在船艙  船票 boat(我也不知道怎麼解釋)  性別

# 篩選特徵值和目標值
x = titanic[["pclass", "age", "sex","embarked","home.dest","room"]]
y = titanic["survived"]
row.names pclass survived name age embarked home.dest room ticket boat sex
0 1 1st 1 Allen, Miss Elisabeth Walton 29.0 Southampton St Louis, MO B-5 24160 L221 2 female
1 2 1st 0 Allison, Miss Helen Loraine 2.0 Southampton Montreal, PQ / Chesterville, ON C26 NaN NaN female
2 3 1st 0 Allison, Mr Hudson Joshua Creighton 30.0 Southampton Montreal, PQ / Chesterville, ON C26 NaN (135) male
'''2 數據預處理'''
x.info() #發現缺失值,
#root變量缺失值太多,不具備解釋性,刪除。
x = x.drop("room", inplace=False, axis=1)   #刪除room列
#age變量缺失值,用均值代替
x["age"] = x["age"].fillna(x["age"].mean(), inplace=False)
#embarked 、home.dest 缺失,使用上一個人的向下填充
x = x.fillna(method="ffill",inplace=False)
#home.dest 缺失值用向上填充
display(x.head(3))

<class ‘pandas.core.frame.DataFrame’>
RangeIndex: 1313 entries, 0 to 1312
Data columns (total 6 columns):
pclass 1313 non-null object
age 633 non-null float64
sex 1313 non-null object
embarked 821 non-null object
home.dest 754 non-null object
room 77 non-null object
dtypes: float64(1), object(5)
memory usage: 61.6+ KB

pclass age sex embarked home.dest
0 1st 29.0 female Southampton St Louis, MO
1 1st 2.0 female Southampton Montreal, PQ / Chesterville, ON
2 1st 30.0 male Southampton Montreal, PQ / Chesterville, ON
x.info() #發現缺失值

<class ‘pandas.core.frame.DataFrame’>
RangeIndex: 1313 entries, 0 to 1312
Data columns (total 5 columns):
pclass 1313 non-null object
age 1313 non-null float64
sex 1313 non-null object
embarked 1313 non-null object
home.dest 1313 non-null object
dtypes: float64(1), object(4)
memory usage: 51.4+ KB

'''3 特徵工程'''
# 1) 轉換成字典
x = x.to_dict(orient="records")
# print(x)

# 2、數據集劃分
from sklearn.model_selection import train_test_split
x_train, x_test, y_train, y_test = train_test_split(x, y, random_state=11)
# display(x_train)

# 3、字典特徵抽取
from sklearn.feature_extraction import DictVectorizer
from sklearn.tree import DecisionTreeClassifier, export_graphviz
transfer = DictVectorizer(sparse=False)
x_train = transfer.fit_transform(x_train)
x_test = transfer.transform(x_test)
# print(transfer.get_feature_names()) #返回類別名稱
# print(x_train)
'''4 模型預估器'''
from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import GridSearchCV
estimator = RandomForestClassifier()
# 加入網格搜索與交叉驗證
# 參數準備
param_dict = {"n_estimators": [100,120,200,300,500,800], "max_depth": [3,5,8,10,15], "max_features":["auto","log2"]}
estimator = GridSearchCV(estimator, param_grid=param_dict, cv=3)
estimator.fit(x_train, y_train)

GridSearchCV(cv=3, error_score=‘raise’,
estimator=RandomForestClassifier(bootstrap=True, class_weight=None, criterion=‘gini’,
max_depth=None, max_features=‘auto’, max_leaf_nodes=None,
min_impurity_decrease=0.0, min_impurity_split=None,
min_samples_leaf=1, min_samples_split=2,
min_weight_fraction_leaf=0.0, n_estimators=10, n_jobs=1,
oob_score=False, random_state=None, verbose=0,
warm_start=False),
fit_params=None, iid=True, n_jobs=1,
param_grid={‘n_estimators’: [100, 120, 200, 300, 500, 800], ‘max_depth’: [3, 5, 8, 10, 15], ‘max_features’: [‘auto’, ‘log2’]},
pre_dispatch=‘2*n_jobs’, refit=True, return_train_score=‘warn’,
scoring=None, verbose=0)

'''5 模型評估'''
# 方法1:直接比對真實值和預測值
y_predict = estimator.predict(x_test)
print("y_predict:\n", y_predict)
#print("直接比對真實值和預測值:\n", y_test == y_predict)

# 方法2:計算準確率
score = estimator.score(x_test, y_test)
print("準確率爲:\n", score)

# 最佳參數:best_params_
print("最佳參數:\n", estimator.best_params_)
# 最佳結果:best_score_
print("最佳結果:\n", estimator.best_score_)
# 最佳估計器:best_estimator_
print("最佳估計器:\n", estimator.best_estimator_)
# 交叉驗證結果:cv_results_
#print("交叉驗證結果:\n", estimator.cv_results_)    transfer.get_feature_names()

y_predict:
[0 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 0 1 1 0 0 0 0 0 0 0
0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 1 0 1 0 0 0 0
1 1 0 1 0 0 0 0 0 1 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 1 1 1 0 0
0 1 0 0 0 1 1 0 1 1 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 1 0 1 1 0 0 1 0 0 0 0
0 0 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 1 0 1 0 0 0 0 0 1 1 1 0 0 0 1 0 1 1 0 0 1 1 0 1 0 1 0 0 0 0 0 0 0
0 0 1 0 0 0 0 0 0 1 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 1
1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0]
準確率爲:
0.8206686930091185
最佳參數:
{‘max_depth’: 15, ‘max_features’: ‘auto’, ‘n_estimators’: 200}
最佳結果:
0.8262195121951219
最佳估計器:
RandomForestClassifier(bootstrap=True, class_weight=None, criterion=‘gini’,
max_depth=15, max_features=‘auto’, max_leaf_nodes=None,
min_impurity_decrease=0.0, min_impurity_split=None,
min_samples_leaf=1, min_samples_split=2,
min_weight_fraction_leaf=0.0, n_estimators=200, n_jobs=1,
oob_score=False, random_state=None, verbose=0,
warm_start=False)

發表評論
所有評論
還沒有人評論,想成為第一個評論的人麼? 請在上方評論欄輸入並且點擊發布.
相關文章