Clever Geek Handbook

Converting probability vector to target vecor in python?

I am doing logistic regression on aperture set from sklearn, I know math and am trying to implement it. In the last step I get a prediction vector, this prediction vector represents the probability that this data point is in class 1 or class 2 (binary classification).

Now I want to turn this prediction vector into a target vector. Let's say if the probability is greater than 50%, then the corresponding data point will belong to class 1, otherwise class 2. Use 0 to represent class 1, 1 for class 2.

I know there is a version for a loop, just iterating over the entire vector. But when the size gets large, it is very expensive for the loop, so I want to do it more efficiently like a matrix operation, it's faster than doing a matrix operation in a loop.

Any suggestion for a faster method?

+4


source to share


4 answers 


import numpy as np

a = np.matrix('0.1 0.82')
print(a)

a[a > 0.5] = 1
a[a <= 0.5] = 0
print(a)

Output:

[[ 0.1   0.82]]
[[ 0.  1.]]


Update:

import numpy as np

a = np.matrix('0.1 0.82')
print(a)

a = np.where(a > 0.5, 1, 0)
print(a)
+7


source


for a multi-class or more generic solution use



np.argmax (y_hat, 1)

0


source


A more general solution for a 2D array that has many vectors with many classes:

import numpy as np
a = np.array( [ [.5, .3, .2], 
                [.1, .2, .7], 
                [ 1,  0,  0] ] )

idx = np.argmax(a, axis=-1)
a = np.zeros( a.shape )
a[ np.arange(a.shape[0]), idx] = 1

print(a)

Output:

[[1. 0. 0.]
 [0. 0. 1.]
 [1. 0. 0.]]
0


source


If you want to continue working with confusion matrix etc. And again get the original format of the target variable in scikit: array([1 0... 1])

you can use:

a = clf.predict_proba(X_test)[:,1]
a = np.where(a>0.5, 1, 0)

[:,1]

refers to the second class (in my case: 1), the first class in my case was 0

0


source






More articles:


All Articles