Question

我一直在努力了解为什么fit_transform不断抛出错误。调试并没有多大帮助，因为它只是将我引向数组的定义，而且我不确定自己缺少什么。有什么想法吗？

import numpy as np
from sklearn.preprocessing import MinMaxScaler
import pickle

def split_data(inputs, outputs, p_train=0.9, p_test=0.1):
    train_size = int(len(inputs) * p_train)
    test_size = int(len(inputs) * p_test)
    inputs_train, inputs_test = inputs[0:train_size, :], inputs[train_size:, :]
    outputs_train, outputs_test = outputs[0:train_size, :], outputs[train_size:, :]
    return inputs_train, inputs_test, outputs_train, outputs_test

rawinputs = pickle.load(open('rawinputs.pck', 'rb'))  
rawoutputs = pickle.load(open('rawoutputs.pck', 'rb'))  

#split
inputs_train, inputs_test, outputs_train, outputs_test = split_data(rawinputs, rawoutputs, p_train=0.90, p_test=0.10)

#normalize
scaler_inputs = MinMaxScaler()
inputs_train_scaled = scaler_inputs.fit_transform(inputs_train)
inputs_test_scaled = scaler_inputs.transform(inputs_test)
outputs_train = np.asmatrix(outputs_train)

使用fit_transform的第一行引发错误：

inputs_train_scaled = scaler_inputs.fit_transform(inputs_train)

例如一排原始输入数据：

['28,7170876207375' '339,050018316624' '0,173448071160097' '158,211319524893']

引发错误：

File "C:\****\venv\lib\site-packages\sklearn\base.py", line 690, in fit_transform return self.fit(X, **fit_params).transform(X)

Answer 1

关键点是inputs_train变量必须是二维数组。

这里有一个简单的例子：

from sklearn.datasets import fetch_olivetti_faces

我们将MinMaxScaler应用于面部数据集。

from sklearn.datasets import fetch_olivetti_faces

olivetti = fetch_olivetti_faces()

X = olivetti.images
y = olivetti.target

X是人脸图像，其中y是每个人脸的标签。

接下来，我们将数据集分为训练集和测试集。

类似于您的split_data方法。

from sklearn.datasets import fetch_olivetti_faces
from sklearn.model_selection import train_test_split

olivetti = fetch_olivetti_faces()

X = olivetti.images
y = olivetti.target

x_train, x_test, y_train, y_test = train_test_split(X, y,
                                                    test_size=0.2,
                                                    random_state=42)

random_state变量将帮助我们在每次运行代码时产生相同的训练和测试集。

现在，我们来看一下x_train变量：

print(x_train.shape)

有320张火车图像，每个图像为64 x 64像素。我们必须重塑形状以使用MinMaxScaler

x_train = x_train.reshape(x_train.shape[0], x_train.shape[1] * x_train.shape[2])

现在新的图像尺寸：

最终代码：

from sklearn.datasets import fetch_olivetti_faces
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import MinMaxScaler

olivetti = fetch_olivetti_faces()

X = olivetti.images
y = olivetti.target

x_train, x_test, y_train, y_test = train_test_split(X,
                                                    y,
                                                    test_size=0.2,
                                                    random_state=42)

x_train = x_train.reshape(x_train.shape[0], x_train.shape[1] * x_train.shape[2])
x_test = x_test.reshape(x_test.shape[0], x_test.shape[1] * x_test.shape[2])

min_max_obj = MinMaxScaler()

inputs_train_scaled = min_max_obj.fit_transform(X=x_train)
inputs_test_scaled = min_max_obj.transform(X=x_test)

为什么fit_transform不断抛出错误？

1 个答案: