1
00:00:01,090 --> 00:00:06,670
In this lecture, we learned how to calculate performance metrics for our models.

2
00:00:08,330 --> 00:00:14,410
We will look at you loose at these places on record and you'll see off auto secrets.

3
00:00:15,380 --> 00:00:18,110
We have already discussed about this in order to relax chips.

4
00:00:18,890 --> 00:00:22,070
We are not going to see all too calculated and Biton.

5
00:00:24,560 --> 00:00:33,530
You can always calculate this values manually using positive groo, negative, false positive from your

6
00:00:33,530 --> 00:00:34,730
confusion matrix.

7
00:00:35,910 --> 00:00:37,350
What do you going to also find it?

8
00:00:37,560 --> 00:00:39,870
Using a single line of code?

9
00:00:42,930 --> 00:00:48,570
So from Escalon Metrics, we will import precisionist score and record they scored.

10
00:00:49,980 --> 00:00:54,730
Now, to calculate preachiness score, we just have to write Bridgit on this scores.

11
00:00:54,890 --> 00:00:59,140
This is the function and then we have to write two arguments.

12
00:00:59,630 --> 00:01:02,690
First one is the crucial value of Y.

13
00:01:03,080 --> 00:01:05,630
And the second one is the predicted value of five.

14
00:01:06,470 --> 00:01:08,090
So we run the score.

15
00:01:09,720 --> 00:01:13,970
The president for our largest sigmoidal is zero point six five three.

16
00:01:15,180 --> 00:01:18,160
Similarly, to calculate the recall, we just have to right.

17
00:01:18,280 --> 00:01:19,900
Recall underscored his score.

18
00:01:20,290 --> 00:01:27,490
And then again, actual value of play and credit value of where you can see the recall for our largest

19
00:01:27,490 --> 00:01:29,550
sigmoidal is zero point six six.

20
00:01:32,460 --> 00:01:40,160
To calculate the U.S., we have to import ardency, underscore, AUC underscore is scored from Escalon

21
00:01:40,160 --> 00:01:40,650
metrics.

22
00:01:42,640 --> 00:01:43,380
When you do that.

23
00:01:44,470 --> 00:01:51,670
And then to calculate the U.S. score, we just have to write our policy, underscore, AUC, underscore,

24
00:01:51,710 --> 00:01:56,020
rescored and then the actual value of Y and the value of Y.

25
00:01:58,390 --> 00:01:59,330
And this.

26
00:02:01,420 --> 00:02:07,120
So the U.S. for what logistic regression model is zero point six eight five.

27
00:02:08,730 --> 00:02:13,540
We will also see you see skorts for Aldea and Gay and then.

28
00:02:14,060 --> 00:02:20,410
And we will use this EOC scores to compare the accuracy of our models in the later part.

29
00:02:20,440 --> 00:02:21,610
Of course.
