如何从随机生成的值中获取四阶多项式并将数组从 (100) 重塑为 (100,4)？

Question

我正在尝试实现多项式的四阶并使用以下指令实现了以下逻辑：

首先，从 scikit-learn 导入 PolynomialFeatures 函数
并用它来生成一个新的 X_4d 数组，该数组具有最高为四阶特征的所有特征
加入高阶多项式特征变换后的形状应该是(100,4)，前5
示例应如下所示

代码：

import numpy as np
from sklearn.linear_model import LinearRegression
from sklearn.preprocessing import PolynomialFeatures

def forth_order(m, a=-3.15, b=1.18, c=3.52, d=3.92, mu=0.0, sigma=0.33):
    new_x = np.random.uniform(low=-1.0, high=1.0, size=(m,)).reshape(100,4)
    new_y =  a*x**3 + b*x**2 + c*x + d + np.random.randn(m)*sigma
    return(new_x, new_y)

print(new_x)
print(new_x.shape)

print(new_y)    
print(new_y.shape)

# to find the 4th order for random generated values
np.random.seed(42)
# call your function to generate the artificial cubic data set
new_x,new_y = make_cubic_dataset(100)

transformer = PolynomialFeatures(degree=4, include_bias=False)
transformer.fit(new_x).reshape(100,4)
x_ = transformer.transform(new_x)
X_4d = np.polyfit(x, y, 3) # fit a degree 4 (cubic) polynomial to the data

print(X_4d)
print(X_4d.shape)

错误：

ValueError: cannot reshape array of size 100 into shape (100,4)

预期结果：

加入高阶多项式特征变换后的形状应该是(100,4)，前5个和samples应该是下面这样

print(X_4d.shape)
>>> (100, 4)

print(X_4d[:5,:])
>>> [[-0.25091976  0.06296073 -0.01579809  0.00396405]
    [ 0.90142861  0.81257354  0.73247704  0.66027576]
    [ 0.46398788  0.21528476  0.09988952  0.04634753]
    [ 0.19731697  0.03893399  0.00768234  0.00151586]
    [-0.68796272  0.4732927  -0.32560773  0.22400598]]

我遇到了解决这个问题的麻烦。

Answer 1

这是代码中的一个小改动，不言自明：

import numpy as np
from sklearn.linear_model import LinearRegression
from sklearn.preprocessing import PolynomialFeatures

def forth_order(m, a=-3.15, b=1.18, c=3.52, d=3.92, mu=0.0, sigma=0.33):
    new_x = np.random.uniform(low=-1.0, high=1.0, size=(m,)).reshape(-1,1)
    new_y =  a*x**3 + b*x**2 + c*x + d + np.random.randn(m)*sigma
    return(new_x, new_y)

np.random.seed(42)

# call your function to generate the artificial cubic data set for forth order
new_x,new_y = forth_order(100)

X_4d = transformer.transform(new_x)
X_4d = PolynomialFeatures(degree=4, include_bias=False).fit_transform(new_x)
transformer.fit(new_x)

print(X_4d.shape)
print(X_4d[:5,:])

输出

print(X_4d.shape)
>>> (100, 4)

print(X_4d[:5,:])
>>> [[-0.25091976  0.06296073 -0.01579809  0.00396405]
    [ 0.90142861  0.81257354  0.73247704  0.66027576]
    [ 0.46398788  0.21528476  0.09988952  0.04634753]
    [ 0.19731697  0.03893399  0.00768234  0.00151586]
    [-0.68796272  0.4732927  -0.32560773  0.22400598]]

如何从随机生成的值中获取四阶多项式并将数组从 (100) 重塑为 (100,4)？

How to get the 4th order of polynomial from random generated values and reshape array from (100) to (100,4)?

python

linear-regression

python-3.x

scikit-learn

sklearn-pandas