Python ML - How to best rescue python ml value array

Question

I created a model with sklearn for classification. When I call the function y_pred2 = clf.predict (features2) it returns a list with all the ids of my prediction

y_pred2 = clf.predict(features2)

Print Array

array([**5**, 5, 5, 1, 6, 1, 6, 1, 1, 1])

Up here I can add it to my dataframe without any problem df ['prediction'] = y_pred2

But I also want to record my multi-class probability at the best value

y_pred2 = clf.predict_proba(features2)[0]

array([0.02670249, 0.23888486, 0.00940765, 0.15213608, 0.02719888,
   **0.42038983**, 0.07503347, 0.02960037, 0.02064636])

But I also want to record the probability of my multi-class by the best value, but the return from the predict_proba function returns an array of all my class, how can I do to record the value of my prediction score.

Example according to my prediction my best class and first position 5 where probability value was 0.42038983. C

How can I take the best value from my array and write to my dataframe?

how did your Y look like before training?

Jeril
– Jeril

2020-02-06 02:41:49 +00:00
Commented Feb 6, 2020 at 2:41 — Jeril
– Jeril, Commented Feb 6, 2020 at 2:41

mohsinali · Accepted Answer · 2020-02-06 04:15:00Z

1

so if we deal with the result as a normal list, we can try something like

y_pred2 = clf.predict_proba(features2)[0]
someclass = y_pred2.index(max(y_pred2)) # returns the class index *Maximum predicted value
itsprob = max(y_pred2) # returns the Portability *Maximum predicted value

then you can either class,prob variable in a df or how ever you like. i cant reproduce your whole scenario as i dont know all of your code

answered Feb 6, 2020 at 4:15

mohsinali

2961 silver badge9 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Python ML - How to best rescue python ml value array

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related