Skip to main content
 首页 » 编程设计

python之来自两个数组的最小二乘法最佳拟合线的斜率 m

2024年11月01日15duanxz

给定两个文本文件和作为数组的“x”和“y”,编写一个子程序,给出最小二乘法最佳拟合线的斜率“m”和截距“b”。

x = 1,2,3,4

y = 5,6,7,8

m = ((sum(xy)-sum(x)ybar)/((sumx^2)-(sum(x)xbar))

b = ybar - m(xbar).

xbar and ybar are the mean values of the x and y arrays.

我的导师说我们不能使用 np.loadtxt 之外的内部 Python 函数。

除了定义 4 个数组组合,然后是 4 个组合中的每一个的总和,然后找到 m 之外,我正在尝试找到一种更有效的方法来执行此操作。

我尝试将数组组合的每个总和定义为嵌套循环。

import numpy as np 
 
x = np.loadtxt("C:/Users/Bob/Documents/School/Comp Phys/xlsf.dat") 
y = np.loadtxt("C:/Users/Bob/Documents/School/Comp Phys/ylsf.dat") 
 
xbar = np.mean(x) 
ybar = np.mean(y) 
 
nx = len(x) 
ny = len(y) 
 
q = print(xbar) 
 
sum1=[[0 for i in range(nx)] for j in range(ny)] 
 
for i in range (0, nx): 
    for j in range (0, ny): 
        sum1[i][j] = x[i] * y[j] 
 
for i in range (ny): 
        for j in range (nx): 
            ((sum1[i][j])) 
 
sum2=[[0 for q in range(nx)] for r in range(ybar)] 
 
for q in range (0, nx): 
    for r in range (0, ybar): 
        sum2[q][r] = x[q] * ybar[r] 
 
for q in range (ybar): 
    for r in range (nx): 
        ((sum2[q][r])) 

Python 正在处理 sum2 部分,出现以下错误:

sum2=[[0 for q in range(nx)] for r in range(ybar)]

TypeError: 'numpy.float64' object cannot be interpreted as an integer.

请您参考如下方法:

xbar = np.mean(x) 
ybar = np.mean(y) 

在您的代码中,xbar 和 ybar 是数组的平均值,以 float 形式给出(numpy 类型 float 64)。

在下面的行中,您使用 range(ybar) 作为 for 循环。

sum2=[[0 for q in range(nx)] for r in range(ybar)] 
 
for q in range (0, nx): 
    for r in range (0, ybar): 
        sum2[q][r] = x[q] * ybar[r] 
 
for q in range (ybar): 
    for r in range (nx): 
        ((sum2[q][r])) 

但是,你的for循环不能使用非整数来循环,因为它们不能循环,比如4.5次。您可能需要构建代码以不同的方式计算总和,例如使用不同的变量作为循环范围。

我仍然不确定您要尝试进行哪些计算,但我已经构建了以下代码,您可以通过这些代码进行计算。它很简单,按原样执行每一步,让您优化代码本身以更好地运行,但希望它能让您了解如何构建计算。

# sum1 array, x*y 
sum1=[[0 for i in range(nx)] for j in range(ny)] 
for i in range (0, nx): 
    for j in range (0, ny): 
        sum1[i][j] = x[i] * y[j] 
 
# sum2 array, x 
# notice that the array sum2 can just be completely replaced by x[] 
sum2=[[0 for q in range(nx)] for r in range(ny)] 
for q in range (0, nx): 
        sum2[q][0] = x[q]  
 
# sum3 array, x^2 
# same for sum3, we can just replace it with x[]*x[] 
sum3 = [[0 for q in range(nx)] for r in range(ny)] 
for q in range (0, nx): 
        sum2[q][0] = x[q]*x[q] 
 
term1 = 0 
term2 = 0 
term3 = 0 
term4 = 0 
 
# sum each 'sum array' and apply them to the term variables 
for i in range(nx): 
    for j in range(ny): 
        term1 += sum1 
        term2 += sum2 
        term3 += sum3 
        term4 += sum2 
 
term2 = term2 * ybar // we multiply that sum by the coefficients  
term4 = term4 * xbar 
 
m = (term1 - term2 ) / (term3 - term4 ) // calculate m 
b = ybar - m * xbar // calculate b