Suppose you have created a classification model. later you have evaluated the model with a confusion matrix learned from class. the table below shows all four values from the test data:
tp: 10 9 8 8 6 5 5 4 3 3 1 0
fp: 10 10 9 8 8 8 6 6 5 2 2 0
fn: 0 1 2 2 4 5 5 6 7 7 9 10
tn: 0 0 1 2 2 2 4 4 5 8 8 10
1. calculate true positive rate for each experiment (column) from above data.
2. calculate false positive rate for each experiment (column) from above data.
3. generate the roc curve from your calculations. you can use any available tool to do so. paste the picture in your submission.
4. does this model, according to the roc curve you have generated, make the prediction better than a random guess? justify your answer.
5. what is the highest value of accuracy (see the lecture and notes for the definition of "accuracy") this model has reached, according to these ten test data?