[英]How to keep the ones in the Diagonal of sns.heatmap?
I want to plot the correlation Matrix with sns.heatmap and have some questions.我想 plot 与 sns.heatmap 的相关矩阵并有一些问题。 This is my code:这是我的代码:
plt.figure(figsize=(8,8)) mask =np.zeros_like(data.corr()) mask[np.triu_indices_from(mask)] = True sns.heatmap(data.corr(), mask=mask, linewidth=1, annot=True, fmt=".2f",cmap='coolwarm',vmin=-1, vmax=1) plt.show()
and this is what i get: [Correlation Matrix][1] [1]: https://i.stack.imgur.com/DX2oN.png \这就是我得到的:[相关矩阵][1] [1]: https://i.stack.imgur.com/DX2oN.png \
Now i have some questions:现在我有一些问题:
1) How can i keep the ones in the diagonale? 1)我怎样才能把那些放在对角线上?
2) How can i change the position of the x-axis? 2)如何改变x轴的position?
3) I want that the colorbar goes from 1 till -1, but the code is not working 3) 我希望颜色条从 1 变为 -1,但代码不工作
I hope someone can help.我希望有人能帮帮忙。
Thx谢谢
I think you have to check data.corr()
, because your code is correct and gives the diagnoal (see below). 我认为您必须检查data.corr()
,因为您的代码正确并且可以诊断(请参阅下文)。 One question is: you use np.triu
but the picture you show displays np.tirl
. 一个问题是:您使用np.triu
但显示的图片显示np.tirl
。
Here the code I've tested - the diagonal is there: 这是我测试过的代码-对角线在这里:
N = 5
A = np.arange(N*N).reshape(N,N)
B = np.tril(A)
mask =np.zeros_like(A)
mask[np.triu_indices_from(mask)] = True
print('A'); print(A); print()
print('tril(A)'); print(B); print()
print('mask'); print(mask); print()
gives 给
A
[[ 0 1 2 3 4]
[ 5 6 7 8 9]
[10 11 12 13 14]
[15 16 17 18 19]
[20 21 22 23 24]]
tril(A)
[[ 0 0 0 0 0]
[ 5 6 0 0 0]
[10 11 12 0 0]
[15 16 17 18 0]
[20 21 22 23 24]]
mask
[[1 1 1 1 1]
[0 1 1 1 1]
[0 0 1 1 1]
[0 0 0 1 1]
[0 0 0 0 1]]
edit: suplement 编辑:补充
you could re-fine the mask, eg 您可以重新调整面膜,例如
C = A *mask
D = np.where(C > 1, 1,C)
print('D'); print(D)
gives 给
D
[[0 1 1 1 1]
[0 1 1 1 1]
[0 0 1 1 1]
[0 0 0 1 1]
[0 0 0 0 1]]
The first element of the diagonal of D is now a Zero since the first element of the diagonal of A is a Zero too. D的对角线的第一个元素现在为零,因为A的对角线的第一个元素也为零。
edit: suplement 2 编辑:补充2
F = np.tril(A,-1)
E = np.eye(N)
G = E + F
print('F'); print(F); print()
print('E'); print(E); print()
print('G'); print(G); print()
gives 给
F
[[ 0 0 0 0 0]
[ 5 0 0 0 0]
[10 11 0 0 0]
[15 16 17 0 0]
[20 21 22 23 0]]
E
[[1. 0. 0. 0. 0.]
[0. 1. 0. 0. 0.]
[0. 0. 1. 0. 0.]
[0. 0. 0. 1. 0.]
[0. 0. 0. 0. 1.]]
G
[[ 1. 0. 0. 0. 0.]
[ 5. 1. 0. 0. 0.]
[10. 11. 1. 0. 0.]
[15. 16. 17. 1. 0.]
[20. 21. 22. 23. 1.]]
mask[np.triu_indices_from(mask)]
will define the triangle (including diagonal) mask[np.triu_indices_from(mask)]
将定义三角形(包括对角线)
mask[np.eye(mask.shape[0], dtype=bool)]
will define the diagonal. mask[np.eye(mask.shape[0], dtype=bool)]
将定义对角线。
If you put those together, you can control them independently.如果将它们放在一起,则可以独立控制它们。 (Be aware you need to set the triangle before the diagonal). (请注意,您需要在对角线之前设置三角形)。
def plot_correlation_matrix(df, remove_diagonal=True, remove_triangle=False, **kwargs):
corr = df.corr()
# Apply mask
mask = np.zeros_like(corr, dtype=np.bool)
mask[np.triu_indices_from(mask)] = remove_triangle
mask[np.eye(mask.shape[0], dtype=bool)] = remove_diagonal
# Plot
# plt.figure(figsize=(8,8))
sns.heatmap(corr, mask=mask, **kwargs)
plt.show()
So this command will generate the matrix, removing the upper triangle, but keeping the diagonal:所以此命令将生成矩阵,移除上三角,但保留对角线:
plot_correlation_matrix(df[colunas_notas], remove_diagonal=False, remove_triangle=True)
Change of the position of the x-axis x轴位置的变化
Since I'm not experienced with seaborn I would use matplotlib to plot the heat map ( here an example ) an then use matplotlib's twinx()
or twiny()
to place the axis where you want to have it ( here an example ). 由于我不seaborn经历我会使用matplotlib绘制热图( 这里的例子 )的再使用matplotlib的twinx()
或twiny()
放置轴,你想拥有它( 这里的例子 )。
(I think that can be done with seaborn too - I just do not know it) (我认为也可以使用seaborn来完成-我只是不知道)
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.