从列表生成邻接矩阵,其中邻接意味着相等的元素

时间:2017-10-28 21:22:25

标签: python numpy combinations itertools adjacency-matrix

我有一个这样的清单:

lst = [0, 1, 0, 5, 0, 1]

我想生成一个邻接矩阵:

out = 
array([[ 1.,  0.,  1.,  0.,  1.,  0.],
       [ 0.,  1.,  0.,  0.,  0.,  1.],
       [ 1.,  0.,  1.,  0.,  1.,  0.],
       [ 0.,  0.,  0.,  1.,  0.,  0.],
       [ 1.,  0.,  1.,  0.,  1.,  0.],
       [ 0.,  1.,  0.,  0.,  0.,  1.]])

其中out[i,j] = 1 if lst[i]==lst[j]

这是我的代码,包含两个for循环:

lst = np.array(lst)
label_lst = list(set(lst))
out = np.eye(lst.size, dtype=np.float32)
for label in label_lst:
  idx = np.where(lst == label)[0]
  for pair in itertools.combinations(idx,2):
    out[pair[0],pair[1]] = 1
    out[pair[1],pair[0]] = 1

但我觉得应该有办法改善这一点。有什么建议吗?

2 个答案:

答案 0 :(得分:3)

使用broadcasted comparison -

np.equal.outer(lst, lst).astype(int) # or convert to float

示例运行 -

In [787]: lst = [0, 1, 0, 5, 0, 1]

In [788]: np.equal.outer(lst, lst).astype(int)
Out[788]: 
array([[1, 0, 1, 0, 1, 0],
       [0, 1, 0, 0, 0, 1],
       [1, 0, 1, 0, 1, 0],
       [0, 0, 0, 1, 0, 0],
       [1, 0, 1, 0, 1, 0],
       [0, 1, 0, 0, 0, 1]])

或转换为数组,然后手动扩展到2D并比较 -

In [793]: a = np.asarray(lst)

In [794]: (a[:,None]==a).astype(int)
Out[794]: 
array([[1, 0, 1, 0, 1, 0],
       [0, 1, 0, 0, 0, 1],
       [1, 0, 1, 0, 1, 0],
       [0, 0, 0, 1, 0, 0],
       [1, 0, 1, 0, 1, 0],
       [0, 1, 0, 0, 0, 1]])

答案 1 :(得分:2)

虽然来自@Divakar的建议非常好,但我会把它留在这里作为一个没有numpy的解决方案。

lst = [0, 1, 0, 5, 0, 1]
print([[1 if x==y else 0 for x in lst ] for y in lst])

对于大型列表,接受的解决方案要快得多。