confusion matrix, precision-recall interpretation - caffe AlexNet binary classification

[Suyog Trivedi]

Hello Everyone,

I am trying to train AlexNet for binary class classification. Input for binary class has data imbalance. That is why i am calculating precision and recall to verify the model performance. Can somebody please interpret the outputs given below and tell me what we can/can't say about the trained model.

Input Data:

Class 0 - 280000 Images  (80% training, 10% validation, 10% testing)

Class 1 - 36000 Images (80% training, 10% validation, 10% testing)

(approx.)

Model:

Alexnet 

Batch Size 128

Input Image size - 64*128 (W*H)

Output parameters:

Confusion matrix:

(r , p) | count

(0 , 0) | 31666

(0 , 1) | 182

(1 , 0) | 5

(1 , 1) | 6709

Class 0 :- 

precision=0.973588738935

recall=0.999255287459

Class 1 :- 

precision=0.999842126867

recall=0.994285355438

Thank You!!!