Python scikit-learn SVM 分类器“ValueError: Found array with dim 3. Expected <= 2" [英] Python scikit-learn SVM Classifier &quot;ValueError: Found array with dim 3. Expected &lt;= 2&quot;

问题描述

我正在尝试在 MNIST 数据集上实现 SVM 分类器.由于我的参数是 3 维的，因此会引发以下错误:

I am trying to implement SVM Classifier over MNIST dataset. As my parameters are 3 dimensional its throwing the following error:

ValueError: Found array with dim 3. Expected <= 2

以下是我的代码片段:

import mnist
from sklearn import svm

training_images, training_labels = mnist.load_mnist("training", digits = [1,2,3,4])
classifier = svm.SVC()
classifier.fit(training_images, training_labels)

sklearn 是否支持多维分类器?

Does sklearn support a multi-dimensional classifier?

推荐答案

问题在于您的输入数据.

The problem is with your input data.

您也可以使用 sklearn 来加载数字数据集:

You can use sklearn to load a digit dataset as well:

from sklearn.datasets import load_digits
from sklearn import svm

digits = load_digits()
X = digits.data
y = digits.target

classifier = svm.SVC()
classifier.fit(X[:1000], y[:1000])
predictions = classifier.predict(X[1000:])

这篇关于Python scikit-learn SVM 分类器“ValueError: Found array with dim 3. Expected <= 2"的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！