happy machine learning(Second One)

发现机器学习就根本停不下来

今天来用RNN算法来爽爽僵尸网络宿主预测

首先我们下载好数据，然后打开我们可爱的熊猫

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from subprocess import check_output

df = pd.read_csv('F:\\machine_learning\\network.csv')
df.head(2)
print(df.head())

下面继续处理数据：

按每天的数据包总量来分

df['date']= pd.to_datetime(df['date'])
df = df.groupby(['date','l_ipn'],as_index=False).sum()
df['yday'] = df['date'].dt.dayofyear
df['wday'] = df['date'].dt.dayofweek

对每个ip进行分类：

ip0 = df[df['l_ipn']==0]
max0 = np.max(ip0['f'])
ip1 = df[df['l_ipn']==1]
max1 = np.max(ip1['f'])
ip2 = df[df['l_ipn']==2]
max2 = np.max(ip2['f'])
ip3 = df[df['l_ipn']==3]
max3 = np.max(ip3['f'])
ip4 = df[df['l_ipn']==4]
max4 = np.max(ip4['f'])
ip5 = df[df['l_ipn']==5]
max5 = np.max(ip5['f'])
ip6 = df[df['l_ipn']==6]
max6 = np.max(ip6['f'])
ip7 = df[df['l_ipn']==7]
max7 = np.max(ip7['f'])
ip8 = df[df['l_ipn']==8]
max8 = np.max(ip8['f'])
ip9 = df[df['l_ipn']==9]
max9 = np.max(ip9['f'])
ip0.head(2)

然后我们输出一下ip0的头

很棒，我们已经成功按ip分类了

然后我们按每年的数据包总量来对每个ip进行图形化计数

首先举个例子：

count, division = np.histogram(ip0['f'],bins=10)
division

他会输出这些

array([   68.,   810.,  1552.,  2294.,  3036.,  3778.,  4520.,  5262.,
        6004.,  6746.,  7488.])

然后我们开始对每个ip下手

f,axarray = plt.subplots(5,2,figsize=(15,20))
count, division = np.histogram(ip0['f'],bins=10)
g = sns.barplot(x=division[0:len(division)-1],y=count,ax=axarray[0,0])
axarray[0,0].set_title("Local IP 0 Flow")

count, division = np.histogram(ip1['f'],bins=10)
sns.barplot(x=division[0:len(division)-1],y=count,ax=axarray[0,1])
axarray[0,1].set_title("Local IP 1 Flow")

count, division = np.histogram(ip2['f'],bins=10)
sns.barplot(x=division[0:len(division)-1],y=count,ax=axarray[1,0])
axarray[1,0].set_title("Local IP 2 Flow")

count, division = np.histogram(ip3['f'],bins=10)
sns.barplot(x=division[0:len(division)-1],y=count,ax=axarray[1,1])
axarray[1,1].set_title("Local IP 3 Flow")

count, division = np.histogram(ip4['f'],bins=10)
sns.barplot(x=division[0:len(division)-1],y=count,ax=axarray[2,0])
axarray[2,1].set_title("Local IP 4 Flow")

count, division = np.histogram(ip5['f'],bins=10)
sns.barplot(x=division[0:len(division)-1],y=count,ax=axarray[2,1])
axarray[2,1].set_title("Local IP 5 Flow")

count, division = np.histogram(ip6['f'],bins=10)
sns.barplot(x=division[0:len(division)-1],y=count,ax=axarray[3,0])
axarray[3,0].set_title("Local IP 6 Flow")

count, division = np.histogram(ip7['f'],bins=10)
sns.barplot(x=division[0:len(division)-1],y=count,ax=axarray[3,1])
axarray[3,1].set_title("Local IP 7 Flow")

count, division = np.histogram(ip8['f'],bins=10)
sns.barplot(x=division[0:len(division)-1],y=count,ax=axarray[4,0])
axarray[4,0].set_title("Local IP 8 Flow")

count, division = np.histogram(ip9['f'],bins=10)
sns.barplot(x=division[0:len(division)-1],y=count,ax=axarray[4,1])
axarray[4,1].set_title("Local IP 9 Flow")

巴特西

happy machine learning(Second One)

最新文章

热门文章