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] == class 1, then in the order: TP,FP,FN,TN
#maybe here is something wring:
for i in alle:
pred=predicted
for k in range(len(predicted)):
if float(i[0]) == y_test[k]:
if float(i[0]) == pred[k]:
i[1]+=1
else:
i[2]+=1
elif pred[k] == float(i[0]):
i[3]+=1
elif pred[k] !=float(i[0]) and y_test[k] !=float(i[0]):
i[4]+=1
#specs looks like this: [1, 71, 51, 103, 208]
sens=specs[1]/float(specs[1]+specs[3])
if I'm calculatig
sens=specs[1]/float(specs[1]+specs[2]) im getting also the recall like in
the matrix.
On 17 June 2015 at 15:42, Andreas Mueller <[email protected]> wrote:
> 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 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 <[email protected]> wrote:
>
>> 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 <[email protected]> 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 getting different values when I'm comparing
>>> my Values with those from the metrics.classification_report function.
>>> The general problem ist, my predicted sensitivity is in the
>>> classification report the precision value. I'm computing every sensitivity
>>> with the one vs all approach. So e.g. Class 1 == true, class 2,3,4,5 are
>>> the rest (not true).
>>>
>>> I did this only to get the specificity, and to compare if i computed
>>> everything right.
>>>
>>>
>>>
>>> ----------- ensemble -----------
>>>
>>> precision recall f1-score support
>>>
>>> 1.0 * 0.56 * 0.68 0.61 129
>>> 2.0 *0.28* 0.15 0.20 78
>>> 3.0 * 0.45 * 0.47 0.46 116
>>> 4.0 *0.29* 0.05 0.09 40
>>> 5.0 * 0.44 * 0.66 0.53 70
>>>
>>> avg / total 0.43 0.47 0.43 433
>>>
>>>
>>> Class: 1
>>> sensitivity:*0.556962025316*
>>> specificity: 0.850909090909
>>> ballanced accuracy: 0.703935558113
>>>
>>> Class: 2
>>> sensitivity:*0.279069767442*
>>> specificity: 0.830769230769
>>> ballanced accuracy: 0.554919499106
>>>
>>> Class: 3
>>> sensitivity*:0.446280991736*
>>> specificity: 0.801282051282
>>> ballanced accuracy: 0.623781521509
>>>
>>> Class: 4
>>> sensitivity:*0.285714285714*
>>> specificity: 0.910798122066
>>> ballanced accuracy: 0.59825620389
>>>
>>> Class: 5
>>> sensitivity:*0.442307692308*
>>> specificity: 0.927051671733
>>> ballanced accuracy: 0.68467968202
>>>
>>>
>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>>
>>> _______________________________________________
>>> Scikit-learn-general mailing list
>>> [email protected]
>>> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>>>
>>>
>>
>>
>> ------------------------------------------------------------------------------
>>
>>
>>
>> _______________________________________________
>> Scikit-learn-general mailing
>> [email protected]https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>>
>> _______________________________________________
>> Scikit-learn-general mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>>
>>
>
>
> ------------------------------------------------------------------------------
>
>
>
> _______________________________________________
> Scikit-learn-general mailing
> [email protected]https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
>
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> Scikit-learn-general mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
>
------------------------------------------------------------------------------
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
