如何通过线性回归拟合负数据？

Question

我想拟合我的数据并提取其斜率。我使用线性回归。我的数据是一组包含负值的时钟偏移值。这是我的代码：

from scipy import stats
import scipy
import matplotlib.pyplot as plt
plt.style.use('ggplot')
x= [1549808191, 1549808192, 1549808196, 1549808201, 1549808202, 1549808206, 1549808207, 1549808214, 1549808215, 1549808221, 1549808226, 1549808267, 1549808272, 1549808290, 1549808304, 1549808315, 1549808324, 1549808332, 1549808355, 1549808395, 1549808396]
y= ['7', '0', '0', '0', '-2', '4', '-3', '2', '0', '-1', '0', '-2', '-1', '-1','2', '-2', '1', '0', '0', '-1', '-2']
print(x)
print(y)
plt.plot(x,y,'o-')
plt.show()
slope, intercept, r_value, p_value, std_err = scipy.stats.linregress(x, y)
print(slope)

但是，它给了我这个错误：

    ret = umr_sum(arr, axis, dtype, out, keepdims)
TypeError: cannot perform reduce with flexible type

那么，如何解决此错误？线性回归是用此类数据提取拟合参数的最佳方法吗？

Answer 1

问题似乎出在scipy.stats.linregress(x, y)处，因为您的y值是字符串，您在其中执行拟合。您可以使用map将它们转换为整数类型，一切按预期进行

# import commands here 
plt.style.use('ggplot')
x= [1549808191, 1549808192, 1549808196, 1549808201, 1549808202, 1549808206, 1549808207, 1549808214, 1549808215, 1549808221, 1549808226, 1549808267, 1549808272, 1549808290, 1549808304, 1549808315, 1549808324, 1549808332, 1549808355, 1549808395, 1549808396]
y= ['7', '0', '0', '0', '-2', '4', '-3', '2', '0', '-1', '0', '-2', '-1', '-1','2', '-2', '1', '0', '0', '-1', '-2']

plt.plot(x,y,'o-')
plt.show()
slope, intercept, r_value, p_value, std_err = scipy.stats.linregress(x, list(map(int, y)))
print("The slope is %s" %slope)

# The slope is -0.009607415773244879

Answer 2

问题与解决方案

---

正如在其他答案中所说的那样，问题在于Y值是字符串。它部分地为您工作，因为 matplotlib 自动将您的Y字符串更改为数字。但是 scipy 库没有。因此，您需要将列表转换为数字。见下文

from scipy import stats
import scipy
import matplotlib.pyplot as plt
plt.style.use('ggplot')
x= [1549808191, 1549808192, 1549808196, 1549808201, 1549808202, 1549808206, 1549808207, 1549808214, 1549808215, 1549808221, 1549808226, 1549808267, 1549808272, 1549808290, 1549808304, 1549808315, 1549808324, 1549808332, 1549808355, 1549808395, 1549808396]
y= ['7', '0', '0', '0', '-2', '4', '-3', '2', '0', '-1', '0', '-2', '-1', '-1','2', '-2', '1', '0', '0', '-1', '-2']
y = [float(i) for i in y]
print(x)
print(y)
plt.plot(x,y,'o-')
plt.show()
slope, intercept, r_value, p_value, std_err = scipy.stats.linregress(x, y)
print(slope)

Answer 3

看起来您的y是一个字符串列表。您需要将y的类型设置为整数或浮点数才能进行回归。

Answer 4

将y更改为数字列表：y = [7, 0, 0, 0, -2, ...]

以这种方式工作

from scipy import stats
import scipy
import matplotlib.pyplot as plt
plt.style.use('ggplot')
x= [1549808191, 1549808192, 1549808196, 1549808201, 1549808202, 1549808206, 1549808207, 1549808214, 1549808215, 1549808221, 1549808226, 1549808267, 1549808272, 1549808290, 1549808304, 1549808315, 1549808324, 1549808332, 1549808355, 1549808395, 1549808396]
y= [1549808191, 1549808192, 1549808196, 1549808201, 1549808202, 1549808206, 1549808207, 1549808214, 1549808215, 1549808221, 1549808226, 1549808267, 1549808272, 1549808290, 1549808304, 1549808315, 1549808324, 1549808332, 1549808355, 1549808395, 1549808396]
print(y)
plt.plot(x,y,'o-')
plt.show()
slope, intercept, r_value, p_value, std_err = scipy.stats.linregress(x, y)
print(slope)

返回 1