2016-10-21 51 views
1

我正在实施一个使用python和scikit学习的svm模型。我已经达到的地方,我选择了,而我的特点,并将它们合并到看起来像这样一个列表中的点:将列表转换为svm输入

[[17, 14, 14, 7, 14, 14, 14, 7, 14, 14, 1], 
[14, 14, 7, 14, 14, 14, 7, 14, 14, 7, 1], 
[14, 7, 14, 14, 14, 7, 14, 14, 7, 14, 1], 
[7, 14, 14, 14, 7, 14, 14, 7, 14, 7, 1], 
[14, 14, 14, 7, 14, 14, 7, 14, 7, 14, 1], 
[14, 14, 7, 14, 14, 7, 14, 7, 14, 7, 1], 
[14, 7, 14, 14, 7, 14, 7, 14, 7, 13, 1], 
[7, 14, 14, 7, 14, 7, 14, 7, 13, 7, 1], 
[14, 14, 7, 14, 7, 14, 7, 13, 7, 14, 1], 
[14, 7, 14, 7, 14, 7, 13, 7, 14, 10, 1], 
[7, 14, 7, 14, 7, 13, 7, 14, 10, 4, 1], 
[14, 7, 14, 7, 13, 7, 14, 10, 4, 13, 1], 
[7, 14, 7, 13, 7, 14, 10, 4, 13, 13, 1], 
[14, 7, 13, 7, 14, 10, 4, 13, 13, 7, 1], 
[7, 13, 7, 14, 10, 4, 13, 13, 7, 13, 1], 
[13, 7, 14, 10, 4, 13, 13, 7, 13, 3, 1], 
[7, 14, 10, 4, 13, 13, 7, 13, 3, 13, 1], 
[14, 10, 4, 13, 13, 7, 13, 3, 13, 13, 1], 
[10, 4, 13, 13, 7, 13, 3, 13, 13, 3, 1], 
[4, 13, 13, 7, 13, 3, 13, 13, 3, 13, 0], 
[13, 13, 7, 13, 3, 13, 13, 3, 13, 13, 0], 
[13, 7, 13, 3, 13, 13, 3, 13, 13, 14, 0]] 

在每个元组的最后一个数字是标签。我试图找到一种方法来创建一个数据集,可以分离数据和目标以建立一个模型。我在文档中找不到类似的东西。将它转回到Dataframe会更容易吗?

谢谢!

+1

'''import numpy as np''' +'''data = np.array(my_list)'''+'''X = data [:,:-1]'''+'''Y = data [:, -1]如果你对当前表单的描述/假设是真的,就足够了。这是基本的numpy东西。考虑关于如何处理数组的numpy-docs教程。 – sascha

回答

3

你的意思是把标签中的特征分开吗?如果是这样,你可以使用numpy。

from sklearn import svm 
import numpy as np 
data = np.asarray(A) 
X = data[:,:-1] 
y = data[:,-1] 
clf = svm.SVC() 
clf.fit(X, y) 

A是原始数据列表。

+0

我觉得应该是数组,我已经修复了答案,谢谢 –