用pandas函数代替下列语句： elif i>len(data2_0_time3)-4: data2_0_time3.loc[i,'speed_391_24']=0 data2_0_time3.loc[i,'speed_391_23']=0 data2_0_time3.loc[i,'speed_391_22']=0 k=0 r=len(data2_0_time3)-1-i for j in range(1,4): if np.isnan(data2_0_time3.iat[i-j,0])==False: data2_0_time3.loc[i,'speed_391_24']=data2_0_time3.iat[i,0]+data2_0_time3.iat[i-j,0] data2_0_time3.loc[i,'speed_391_23']=data2_0_time3.iat[i,1]+data2_0_time3.iat[i-j,1] data2_0_time3.loc[i,'speed_391_22']=data2_0_time3.iat[i,2]+data2_0_time3.iat[i-j,2] k=k+1 for j in range(1,r): if np.isnan(data2_0_time3.iat[i+j,0])==False: data2_0_time3.loc[i,'speed_391_24']=data2_0_time3.iat[i,0]+data2_0_time3.iat[i+j,0] data2_0_time3.loc[i,'speed_391_23']=data2_0_time3.iat[i,1]+data2_0_time3.iat[i+j,1] data2_0_time3.loc[i,'speed_391_22']=data2_0_time3.iat[i,2]+data2_0_time3.iat[i+j,2] k=k+1 data2_0_time3.loc[i,'speed_391_24']=data2_0_time3.iat[i,0]/k data2_0_time3.loc[i,'speed_391_23']=data2_0_time3.iat[i,1]/k data2_0_time3.loc[i,'speed_391_22']=data2_0_time3.iat[i,2]/k

使用pandas的rolling函数可以代替原始代码中的for循环，通过计算指定区间内的均值来完成相同的功能，具体代码如下：

data2_0_time3['speed_391_24'] = data2_0_time3.iloc[:,0].rolling(window=4, min_periods=1).mean()
data2_0_time3['speed_391_23'] = data2_0_time3.iloc[:,1].rolling(window=4, min_periods=1).mean()
data2_0_time3['speed_391_22'] = data2_0_time3.iloc[:,2].rolling(window=4, min_periods=1).mean()

其中，rolling函数中的参数window指定了滑动窗口的大小，min_periods指定了有效数据的最小数目。由于rolling的均值计算默认是包括当前值的，所以在数据较短时需要指定min_periods参数。

相关问题

import math import matplotlib.pyplot as plt import pandas as pd file_path = r'C:\Users\520605\Desktop\data.xlsx' raw_data = pd.read_excel(file_path, header=0) data = raw_data.values viscosity = data[0:,:2] print(viscosity) density = data[0:,:3] print(density) num_1 =data[-1,0:0] C=torch.zeros(10) def dis(a,b): return math.sqrt((a[0]-b[0])*(a[0]-b[0])+(a[1]-b[1])*(a[1]-b[1])) K = 4 CentPoint = [] for i in range(K): CentPoint.append([torch.randint(0,10000,(1,)).item(), torch.randint(0,10,(1,)).item()]) print(CentPoint) for p in range(10): NewPoint = [[0, 0] for i in range(K)] for i in range(len(viscosity)): mDis=1e9 mC=0 for j in range(len(CentPoint)): cp=CentPoint[j] D = dis([viscosity[1,i].int(), density[1,i].int()], cp) if mDis>D: mDis=D mC=j C[i]=mC NewPoint[mC][0]+=viscosity[1,i] NewPoint[mC][1]+=density[1,i] for i in range(K): CentPoint[i][0]=NewPoint[i][0]/num_1 CentPoint[i][1]=NewPoint[i][1]/num_1 print(CentPoint) cc=list(C) for i in range(len(viscosity)): if cc[i]==0: plt.plot(viscosity[1,i].item(), density[1,i].item(), 'r.') elif cc[i]==1: plt.plot(viscosity[1,i].item(), density[1,i].item(), 'g.') elif cc[i]==2: plt.plot(viscosity[1,i].item(), density[1,i].item(), 'b.') elif cc[i]==3: plt.plot(viscosity[1,i].item(), density[1,i].item(), color='pink', marker='.') elif cc[i]==4: plt.plot(viscosity[1,i].item(), density[1,i].item(), color='orange', marker='.') for CP in CentPoint: plt.plot(CP[0], CP[1], color='black', marker='X') plt.show()

好的，以下是对你的代码进行改进的建议：

1. 在代码开头，需要将导入模块的语句分行书写，以提高代码的可读性；
2. 在读取 Excel 表格时，建议使用相对路径，以使代码在其他环境中也能正常运行；
3. 在计算样本点到聚类中心的距离时，建议使用 NumPy 库提供的矩阵计算方法，以提高代码的效率；
4. 在绘制数据图形时，可以使用 Matplotlib 的颜色映射函数 `plt.cm`，以简化代码。

改进后的代码如下：

import math
import matplotlib.pyplot as plt
import pandas as pd
import torch
import numpy as np

file_path = './data.xlsx'

raw_data = pd.read_excel(file_path, header=0)
data = raw_data.values
viscosity = data[:, :2]
density = data[:, :3]
num_1 = data[-1, 0]

C = torch.zeros(len(viscosity))

def dis(a, b):
    return np.linalg.norm(a-b)

K = 4
CentPoint = []
for i in range(K):
    CentPoint.append([torch.randint(0, 10000, (1,)).item(),
                      torch.randint(0, 10, (1,)).item()])

for p in range(10):
    NewPoint = [[0, 0] for i in range(K)]
    for i in range(len(viscosity)):
        mDis = 1e9
        mC = 0
        for j in range(len(CentPoint)):
            cp = CentPoint[j]
            D = dis([viscosity[i, 1].item(), density[i, 1].item()], cp)
            if mDis > D:
                mDis = D
                mC = j
        C[i] = mC
        NewPoint[mC][0] += viscosity[i, 1].item()
        NewPoint[mC][1] += density[i, 1].item()

    for i in range(K):
        CentPoint[i][0] = NewPoint[i][0] / num_1
        CentPoint[i][1] = NewPoint[i][1] / num_1

cmap = plt.cm.get_cmap('viridis', K)
for i in range(len(viscosity)):
    plt.plot(viscosity[i, 1].item(), density[i, 1].item(), '.', color=cmap(int(C[i])))

for CP in CentPoint:
    plt.plot(CP[0], CP[1], color='black', marker='X')

plt.show()

这段代码中，我们使用了 NumPy 库提供的 `linalg.norm()` 函数计算样本点到聚类中心的距离，使用了 Matplotlib 的 `cm.get_cmap()` 函数获取颜色映射，以简化绘图代码。