朴素贝叶斯

#朴素贝叶斯：
def getdata():
    file = open("E://data.txt")
    filetest = open("E://test.txt")
    line = file.readlines()
    linetest = filetest.readlines()
    data = []
    datatest = []
    for perline in line:
        templine = perline.strip().split(",")
        data.append(templine)
    for perline in linetest:
        templine = perline.strip().split(",")
        datatest.append(templine)
    label = ['unacc','acc','good','vgood']
    return datatest,data,label

#划分数据集：
def classify_data(data):
    unacc = []
    acc = []
    good = []
    vgood =[]
    for perline in data:
        if perline[-1] == 'unacc':
            unacc.append(perline)
        elif perline[-1] == 'acc':
            acc.append(perline)
        elif perline[-1] == 'good':
            good.append(perline)
        else:
            vgood.append(perline)
    return unacc,acc,good,vgood

#获取先验概率和条件概率：
def first_probability(data):
    unacc, acc, good, vgood = classify_data(data)
    unacc_value = len(unacc) / len(data)
    acc_value = len(acc) / len(data)
    good_value = len(good) / len(data)
    vgood_value = len(vgood) / len(data)
    return unacc_value,acc_value,good_value,vgood_value

#获取条件概率：
#tempdata取值为unacc or acc or good or vgood,nature_i是属性
def conditions_probability(tempdata,nature0,nature1,nature2,nature3,nature4,nature5):
    buying = []
    maint = []
    doors = []
    persons = []
    lug_boot = []
    safety = []
    for perline in tempdata:
        buying.append(perline[0])
        maint.append(perline[1])
        doors.append(perline[2])
        persons.append(perline[3])
        lug_boot.append(perline[4])
        safety.append(perline[5])
    buying_value = buying.count(nature0) / len(tempdata)
    #print(buying.count(nature0))
    maint_value = maint.count(nature1) / len(tempdata)
    doors_value = doors.count(nature2) / len(tempdata)
    persons_value = persons.count(nature3) / len(tempdata)
    lug_boot_value = lug_boot.count(nature4) / len(tempdata)
    safety_value = safety.count(nature5) / len(tempdata)
    return buying_value,maint_value,doors_value,persons_value,lug_boot_value,safety_value

#预测：
def predicition(nature0,nature1,nature2,nature3,nature4,nature5):
    probability = []
    datatest, data, label = getdata()
    unacc, acc, good, vgood = classify_data(data)
    unacc_value, acc_value, good_value, vgood_value = first_probability(data)
    newdata = [unacc,acc,good,vgood]
    value = [unacc_value, acc_value, good_value, vgood_value]
    j = 0
    for tempdata in newdata:
        buying_value, maint_value, doors_value, persons_value, lug_boot_value, safety_value = conditions_probability(tempdata,nature0,nature1,nature2,nature3,nature4,nature5)
        probability.append(value[j] * buying_value * maint_value * doors_value * persons_value * lug_boot_value * safety_value)
        j = j+1
    print(label[probability.index(max(probability))])

#计算准确率率：
def accuracy():
    datatest, data, label = getdata()
    unacc, acc, good, vgood = classify_data(data)
    unacc_value, acc_value, good_value, vgood_value = first_probability(data)
    newdata = [unacc,acc,good,vgood]
    value = [unacc_value, acc_value, good_value, vgood_value]
    sum = 0
    for perline in datatest:
        j = 0
        probability = []
        for tempdata in newdata:
            buying_value, maint_value, doors_value, persons_value, lug_boot_value, safety_value = conditions_probability(tempdata,perline[0],perline[1],perline[2],perline[3],perline[4],perline[5])
            probability.append(value[j] * buying_value * maint_value * doors_value * persons_value * lug_boot_value * safety_value)
            j = j+1
        if label[probability.index(max(probability))] == perline[6]:
            sum = sum+1
    print("准确率是:")
    print(sum / len(datatest))
if __name__ == "__main__":
    accuracy()
    predicition('low','low','5more','more','big','med')
    predicition('vhigh','vhigh','2','2','big','high')