sklearn SVM fit（）“值错误：用序列设置数组元素”

Question

我正在使用sklearn在我自己的一组图像上应用支持向量机。图像被放入数据帧中。 我给fit函数传递一个numpy数组，它有2D列表，这些2D列表表示图像，传递给函数的第二个输入是目标列表（目标是数字）。 我总是得到这个错误“ValueError：用序列设置数组元素”。

trainingImages = images.ix[images.partID <=9]
trainingTargets = images.clustNo.ix[images.partID<=9]
trainingImages.reset_index(inplace=True,drop=True)
trainingTargets.reset_index(inplace=True,drop=True)

classifier = svm.SVC(gamma=0.001)
classifier.fit(trainingImages.image.values,trainingTargets.values.tolist())

错误：

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-43-5336fbeca868> in <module>()
      8 classifier = svm.SVC(gamma=0.001)
      9 
---> 10 classifier.fit(trainingImages.image.values,trainingTargets.values.tolist())
     11 
     12 #classifier.fit(t, list(range(0,2899)))

/Library/Frameworks/Python.framework/Versions/3.5/lib/python3.5/site-packages/sklearn/svm/base.py in fit(self, X, y, sample_weight)
    148         self._sparse = sparse and not callable(self.kernel)
    149 
--> 150         X = check_array(X, accept_sparse='csr', dtype=np.float64, order='C')
    151         y = self._validate_targets(y)
    152 

/Library/Frameworks/Python.framework/Versions/3.5/lib/python3.5/site-packages/sklearn/utils/validation.py in check_array(array, accept_sparse, dtype, order, copy, force_all_finite, ensure_2d, allow_nd, ensure_min_samples, ensure_min_features, warn_on_dtype, estimator)
    371                                       force_all_finite)
    372     else:
--> 373         array = np.array(array, dtype=dtype, order=order, copy=copy)
    374 
    375         if ensure_2d:

ValueError: setting an array element with a sequence.

Answer 1

如果您确定维度是正确的，下面的代码/工作流可能会有帮助

import skimage.io as skio
import matplotlib.pyplot as plt
import numpy as np
from sklearn import svm
from sklearn.metrics import accuracy_score
from sklearn.metrics import precision_score
%matplotlib inline

# Load the data
trainingImages = skio.imread_collection('train/images/*.jpg',conserve_memory=True)

# cast to numpy arrays
trainingImages = np.asarray(trainingImages)

# reshape img array to vector
def reshape_image(img):
    return np.reshape(img,len(img)*len(img[0]))

img_reshape = np.zeros((len(trainingImages),len(trainingImages[0])*len(trainingImages[0][0])))

for i in range(0,len(trainingImages)):
    img_reshape[i] = reshape_image(trainingImages[i])

# SVM
clf = svm.SVC(gamma=0.01,C=10,kernel='poly')
clf.fit(img_reshape,trainingTargets.values.tolist())

Answer 2

我也有同样的错误，这是两种可能性之一：

1- Data and labels are not in the same length.

2- For a specific feature vector, the number of elements are not equal.

Answer 3

这可能是因为“trainingImages.image.values”的数组中元素的数目不同。在stackoverflow中检查类似的问题：

ValueError: setting an array element with a sequence. while using SVM in scikit-learn