import pandas as pd
import math
df = pd.read_table("banana.dat", sep=',', skiprows=3)
df.columns = ['At1', 'At2', 'Class'] # 导入数据并添加列名
PosNum = 0 # 样本中类别1的数量
NegNum = 0 # 样本中类别2的数量
PosList = []
NegList = []
# 将样本中两类分类并计数
for i in range(len(df)):
if df['Class'][i] == 1.0:
PosNum += 1
PosList.append(df['At1'][i])
elif df['Class'][i] == -1.0:
NegNum += 1
NegList.append(df['At1'][i])
# 计算先验概率:#F表示先验概率Former
PosPro_F = PosNum / (PosNum + NegNum)
NegPro_F = NegNum / (PosNum + NegNum)
# 建立正态窗函数,输入分别为数据列表,窗宽h,输入量h
def parzen(DataList, h, x):
sum = 0
hn = h
for i in range(len(DataList)):
sum += math.exp(-(((x - DataList[i]) / hn) ** 2) / 2) / (math.sqrt(2 * math.pi) * hn)
return sum / len(DataList)
testnum = ''
while testnum != exit:
testnum = float(input('请输入-3.09-3.19的测试样本值,输入exit退出'))
if testnum == exit:
break
else:
# 实验数据值的类条件概率密度的取值:N表示Number
PosPro
07-06
2156
