python解决数据预处理：将KDDCPU99数据格式转换成libsvm可读的格式

       最近在研究基于支持向量机的入侵检测，使用的是KDDCPU99的数据，但是KDDCPU99数据的格式是cvs格式的，我们用的支持向量机的插件用的是libsvm库，但是libsvm读数据的格式是 label index：属性 index：属性；label是分类的标记，后面是跟着的属性。线面本人用python语言写了这段数据转换代码与大家分享。

有啥问题请留言，可以相互交流的。

import string,csv

protocol_type = []
service = []
flag = []
attack_type = []
dos_type = ['back','land','neptune','pod','smurf','teardrop']
probe_type = ['ipsweep','nmap','portsweep','satan']
r2l_type = ['ftp_write','guess_passwd','imap','multihop','phf','spy','warezclient','warezmaster']
u2r_type = ['buffer_overflow','loadmodule','perl','rootkit']

def main():
        print "data dealing..."
        writer=csv.writer(open('/home/tool/svmdata/kddcup.data_10_percent1.csv','wb'),delimiter=" ")
        reader=csv.reader(open('/home/tool/svmdata/kddcup.data_10_percent.csv','rb'))
        for line in reader:
                #protocol
                if line[1] in protocol_type:
                        pass   #not do deal
                else:
                        protocol_type.append(line[1])
                protocol_index = protocol_type.index(line[1])+1
                line[1]=protocol_index
                #service
                if line[2] in service:
                        pass #not do deal
                else:
                        service.append(line[2])
                service_index = service.index(line[2])+1
                line[2] = service_index
                #flag
                if line[3] in flag:
                        pass #not do deal
                else:
                        flag.append(line[3])
                flag_index = flag.index(line[3])+1
                line[3] = flag_index
                #type
                line[41] = ''.join(line[41].split('.'))
                if line[41] == 'normal':
                        line[41] = 1
                elif line[41] in dos_type:
                        line[41] = 2
                elif line[41] in probe_type:
                        line[41] = 3
                elif line[41] in r2l_type:
                        line[41] = 4
                elif line[41] in u2r_type:
                        line[41] = 5

                temp = line[41]
                line.insert(0,temp)
                del line[42]
                for k in range(1,len(line)):
                        k2str = str(k) + ':'
                        line[k] = k2str + str(line[k])
                writer.writerow(line)

if __name__=="__main__":
        main()