00:01
To solve this question for part a, we need to find the classification matrix for sample.
00:07
So here we will classify the model as if the model predict fraud, when actually is a fraud.
00:23
So here in the matrix we will write actual and secondly the row will be.
00:32
And search records for first event is 30010.
00:37
So we will make three column actual predicted and then records.
00:50
So actually made it fraud.
00:54
First value will be fraud.
00:56
Second predicted is fraud and records are 1310.
01:03
For the second row, it is again fraud.
01:08
Then next is non -fraud and total recorded are 90.
01:16
Third row is non -fraud.
01:22
Then we have fraud for predicted.
01:25
And total are 130.
01:30
Last row is non -fraud.
01:35
And here we have non -fraud 2 in the predicted column.
01:43
Total are 270.
01:46
So here we are done with the model matrix.
01:50
And now the confusing matrix is as for the question, the action and predicted for frauds, non -frauds.
02:04
We need to construct another table here.
02:11
So here the value for fraud and then non -fraud and then we will compute here total.
02:22
So frauds are 310, 130 are non -fraud, then non -frauds are 90 and 270 are non -fraud.
02:32
So total for the row is 400 here.
02:35
Again for the second row total is 400 and now we will find our column total also.
02:47
Column total is 440 here...