[人工智能]python中孤立森林算法实例

python中孤立森林算法实例

使用python中sklearn库自带的IsolationForest构建孤立森林，并训练预测数据，同时使用plt画图展示

import numpy as np
import matplotlib.pyplot as plt
from sklearn.ensemble import IsolationForest
import csv

#读入数据
def loadData(filename):
    data = open(filename,'r',encoding='utf-8')
    reader = csv.reader(data)
    header = next(reader)
    dataset = []
    price = []
    amount = []
    for row in reader:
        dataset.append([float(row[3]),float(row[4])])
        price.append(float(row[3]))
        amount.append(float(row[4]))

    return dataset,price,amount

def iForest(dataset,price,amount):
    clf = IsolationForest()
    ans = clf.fit_predict(dataset)

    price_abnormal = []
    amount_abnormal = []
    price_normal = []
    amount_normal = []

	#将运算得到的异常数据记录保存
    for d in range(0,len(ans)):
        if ans[d] == -1:
            price_abnormal.append(dataset[d][0])
            amount_abnormal.append(dataset[d][1])
        else:
            price_normal.append(dataset[d][0])
            amount_normal.append(dataset[d][1])


    print(price_normal)
    
    #画图展示
    plt.title("IsolationForest")

	#正常数据点
    b1 = plt.scatter(price_normal, amount_normal, c='white',
                     s=20, edgecolor='k')
	#异常数据点
    c = plt.scatter(price_abnormal, amount_abnormal, c='red',
                    s=20, edgecolor='k')
    plt.axis('tight')
    #x轴、y轴的坐标范围
    plt.xlim((0,3000))
    plt.ylim(0,3500 )
    plt.xlabel('Price')
    plt.ylabel('Amount')
    plt.legend([b1,c],['normal points', 'abnormal points'],
           loc="upper left")
    plt.show()

filename = 'lv3测试集.csv'
dataset,price,amount=loadData(filename)
iForest(dataset,price,amount)

效果如图：