Sorry for the late email,
just wanted to say thank you for the grate description!
On 17 June 2015 at 18:29, Joel Nothman joel.noth...@gmail.com wrote:
Scikit-learn has had a default of a weighted (micro-)average. This is a
bit non-standard, so from now users are expected to specify the
To me, those numbers appear identical at 2 decimal places.
On 17 June 2015 at 23:04, Herbert Schulz hrbrt@gmail.com wrote:
Hello everyone,
i wrote a function to calculate the sensitivity,specificity, ballance
accuracy and accuracy from a confusion matrix.
Now i have a Problem, I'm
Yeah that is the rounding of using %2f in the classification report.
On 06/17/2015 09:20 AM, Joel Nothman wrote:
To me, those numbers appear identical at 2 decimal places.
On 17 June 2015 at 23:04, Herbert Schulz hrbrt@gmail.com
mailto:hrbrt@gmail.com wrote:
Hello everyone,
Yeah i know, thats why I'm asking. i thought precision is not the same like
recall/sensitivity.
recall == sensitivity!?
But in this matrix, the precision is my calculated sensitivity, or is the
precision in this case the sensitivity?
On 17 June 2015 at 15:29, Andreas Mueller t3k...@gmail.com
Hm, the sensitivity (TP/[TP+FN]) should be equal to recall, not the
precision. Maybe it would help if you could print the confusion matrices for
a simpler binary case to track what's going on here
On Jun 17, 2015, at 9:32 AM, Herbert Schulz hrbrt@gmail.com wrote:
Yeah i know, thats why
Hello everyone,
i wrote a function to calculate the sensitivity,specificity, ballance
accuracy and accuracy from a confusion matrix.
Now i have a Problem, I'm getting different values when I'm comparing my
Values with those from the metrics.classification_report function.
The general problem
Sensitivity is recall:
https://en.wikipedia.org/wiki/Sensitivity_and_specificity
Recall is TP / (TP + FN) and precision is TP / (TP + FP).
What did you compute?
On 06/17/2015 09:32 AM, Herbert Schulz wrote:
Yeah i know, thats why I'm asking. i thought precision is not the same
like
I actually computed it like this, maybe I did something in my TP,FP,FN,TN
calculation wrong?
c1,c2,c3,c4,c5=[1,0,0,0,0],[2,0,0,0,0],[3,0,0,0,0],[4,0,0,0,0],[5,0,0,0,0]
alle=[c1,c2,c3,c4,c5]
#as i mentioned 1 vs all, so the first element in the array is just the
class
#[1,0,0,0,0] ==
Ok i think i have it, thanks everyone for the help!
But there is an another problem.
How are you calculating the avg?
example:
--- k-NN ---
precisionrecall f1-score support
1.0 0.50 0.43 0.46 129
2.0 0.31
Scikit-learn has had a default of a weighted (micro-)average. This is a bit
non-standard, so from now users are expected to specify the average when
using precision/recall/fscore. Once
https://github.com/scikit-learn/scikit-learn/pull/4622 is merged,
classification_report will show all the common
About the average: The two common scenarios are micro and macro average (I
think macro is typically the default in scikit-learn) -- you calculated the
macro average in your example.
To further explain the difference betw. macro and micro, let's consider a
simple 2-class scenario and calculate
11 matches
Mail list logo
