快捷導(dǎo)航

python如何去除異常值和缺失值的插值

更新時間：2022年01月27日 10:55:52 作者：Word_gebei

大家好，本篇文章主要講的是python如何去除異常值和缺失值的插值，感興趣的同學(xué)趕快來看一看吧，對你有幫助的話記得收藏一下

1.使用箱型法去除異常值：

import numpy as np
import pandas as pd
import matplotlib as plt
import os
data = pd.read_excel('try.xls', header=0)
# print(data.shape)
# print(data.head(10))
# print(data.describe())
neg_list = ['位移']
print("（1）數(shù)據(jù)的行數(shù)為：")
R = data.shape[0]
print(R)
print("（2）小于或大于閾值的數(shù)據(jù)提?。?)
for item in neg_list:
    neg_item = data[item]<2000
    print(item + '小于2000的有' + str(neg_item.sum()) + '個')
print("（3）異常值的個數(shù)：")
for item in neg_list:
    iqr = data[item].quantile(0.75) - data[item].quantile(0.25)
    q_abnormal_L = data[item] < data[item].quantile(0.25) - 1.5 * iqr
    q_abnormal_U = data[item] > data[item].quantile(0.75) + 1.5 * iqr
    print(item + '中有' + str(q_abnormal_L.sum() + q_abnormal_U.sum()) + '個異常值')
print("（4）箱型圖確定上下限：")
for item in neg_list:
    iqr = data[item].quantile(0.75) - data[item].quantile(0.25)
    Too_small = data[item].quantile(0.25) - 1.5 * iqr
    Too_big = data[item].quantile(0.25) + 1.5 * iqr
print("下限是", Too_small)
print("上限是", Too_big )
print("（5）所有數(shù)據(jù)為：")
a = []
for i in neg_list:
    a.append(data[i])
print(a)
print("（6）所有正常數(shù)據(jù)：")
b = []
j = 0
while j < R:
    if (a[0][j] > Too_small):
        if (a[0][j] < Too_big):
            b.append(a[0][j])
    j += 1
print(b)
print("（7）所有異常數(shù)據(jù)：")
c = []
i = 0
while i < R:
    if (a[0][i] < Too_small or a[0][i] > Too_big):
        c.append(a[0][i])
        a[0][i] = None
    i +=1
print(c)
print("（8）把所有異常數(shù)據(jù)刪除后：")
print(a)
print("（9）所有數(shù)據(jù)處理后輸出：")
d = []
k = 0
while k < R:
    d.append(a[0][k])
    k +=1
print(d)
df = pd.DataFrame(d,columns= ['位移'])
df.to_excel("try_result.xls")

2.拉格朗日插值：

import os
import pandas as pd
import numpy as np
from scipy.interpolate import lagrange
import matplotlib.pyplot as plt

plt.rcParams['font.sans-serif']=['SimHei'] #用來正常顯示中文標(biāo)簽

plt.rcParams['axes.unicode_minus']=False #用來正常顯示負(fù)號

# 數(shù)據(jù)的讀取
data = pd.read_excel('try.xls', header=0)
neg_list = ['位移']

# 數(shù)據(jù)的行數(shù)
R = data.shape[0]

# 異常數(shù)據(jù)的個數(shù)
for item in neg_list:
    iqr = data[item].quantile(0.75) - data[item].quantile(0.25)
    q_abnormal_L = data[item] < data[item].quantile(0.25) - 1.5 * iqr
    q_abnormal_U = data[item] > data[item].quantile(0.75) + 1.5 * iqr
    # print(item + '中有' + str(q_abnormal_L.sum() + q_abnormal_U.sum()) + '個異常值')

# 確定數(shù)據(jù)上限和下限
for item in neg_list:
    iqr = data[item].quantile(0.75) - data[item].quantile(0.25)
    Too_small = data[item].quantile(0.25) - 1.5 * iqr
    Too_big = data[item].quantile(0.25) + 1.5 * iqr

data[u'位移'][(data[u'位移']<Too_small) | (data[u'位移']>Too_big)] = None #過濾異常值，將其變?yōu)榭罩?

#s為列向量，n為被插值位置，k為取前后的數(shù)據(jù)個數(shù)
def ployinter(s,n,k=5):
    y = s[list(range(n-k,n)) + list(range(n+1,n+1+k))]
    y = y[y.notnull()] #剔除空值
    return lagrange(y.index,list(y))(n)

#逐個元素判斷是否需要插值
for i in data.columns:
    for j in range(len(data)):
        if(data[i].isnull())[j]:
            data[i][j] = ployinter(data[i],j)
# print(data[u'位移'])

# 輸出拉格朗日插值后的數(shù)據(jù)
data.to_excel("try_result.xls")

# 把表格列數(shù)據(jù)調(diào)整為arr,arr為修改后的數(shù)據(jù)
print("拉格朗日插值后的數(shù)據(jù)：")
d = []
k = 0
while k < R:
    d.append(data[u'位移'][k])
    k +=1
# print(d)
arr = np.array(d)
print(arr)

# 輸出圖像
x = np.arange(len(d))

plt.plot(x,d,'b-',label="one", marker='*',markersize=4,linewidth=1) # b代表blue顏色  -代表直線

plt.title('位移曲線')

plt.legend(loc='upper left',bbox_to_anchor=(1.0,1.0))

# 直接更改X軸坐標(biāo)數(shù)
# plt.xticks((0,1,2,3,4,5,6,7,8),('0', '1', '2', '3', '4', '5', '6', '7', '8'))

plt.xlabel('時間/h')
plt.ylabel('位移/mm')

#plt.grid(x1)
plt.show()

3.數(shù)據(jù)擬合：

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from scipy.optimize import leastsq


def Fun(p, x):  # 定義擬合函數(shù)形式
    a1, a2, a3 , a4 = p
    return a1 * x ** 3 + a2 * x ** 2 + a3 * x + a4


def error(p, x, y):  # 擬合殘差
    return Fun(p, x) - y


def main():
    x = np.linspace(1, 31, 31)  # 創(chuàng)建時間序列
    data = pd.read_excel('try.xls', header=0)
    y = data[u'位移']
    p0 = [0.1, -0.01, 100, 1000]  # 擬合的初始參數(shù)設(shè)置
    para = leastsq(error, p0, args=(x, y))  # 進(jìn)行擬合
    y_fitted = Fun(para[0], x)  # 畫出擬合后的曲線

    plt.figure
    plt.plot(x, y, 'r', label='Original curve')
    plt.plot(x, y_fitted, '-b', label='Fitted curve')
    plt.legend()
    plt.show()
    print(para[0])


if __name__ == '__main__':
    main()

4.輸出圖像：

import pandas as pd

import numpy as np

import matplotlib.pyplot as plt

plt.rcParams['font.sans-serif']=['SimHei'] #用來正常顯示中文標(biāo)簽

plt.rcParams['axes.unicode_minus']=False #用來正常顯示負(fù)號

jiaodu = ['0', '15', '30', '15', '60', '75', '90', '105', '120']

x = range(len(jiaodu))

y = [85.6801,   7.64586,    86.0956,    159.229,    179.534,    163.238,    96.4436,    10.1619,    90.9262,]

#plt.figure(figsize=(10, 6))

plt.plot(x,y,'b-',label="1", marker='*',markersize=7,linewidth=3) # b代表blue顏色  -代表直線

plt.title('各個區(qū)域亮度變化')

plt.legend(loc='upper left',bbox_to_anchor=(1.0,1.0))

plt.xticks((0,1,2,3,4,5,6,7,8),('0', '15', '30', '15', '60', '75', '90', '105', '120'))

plt.xlabel('角度')

plt.ylabel('亮度')

#plt.grid(x1)

plt.show()

到此這篇關(guān)于python如何去除異常值和缺失值的插值的文章就介紹到這了,更多相關(guān)python去除異常值和缺失值內(nèi)容請搜索腳本之家以前的文章或繼續(xù)瀏覽下面的相關(guān)文章希望大家以后多多支持腳本之家！

您可能感興趣的文章: