1
00:00:00,380 --> 00:00:08,920
Or Rado Rado rat modelling Part 3 churning once you're happy with your models initial performance on

2
00:00:08,920 --> 00:00:10,570
your training dataset.

3
00:00:10,600 --> 00:00:15,010
The next step is to try and improve it like what we said in last lesson.

4
00:00:15,010 --> 00:00:21,850
How a car can be tuned for different styles of driving a model can be tuned for different types of data.

5
00:00:21,850 --> 00:00:28,050
Specifically your data usually this training will take place on a validation data split.

6
00:00:28,380 --> 00:00:34,680
However if you don't have access to a validation set due to how you've done your your training validation

7
00:00:34,680 --> 00:00:41,610
and test data split it can also happen on the training data on many models have different hyper parameters

8
00:00:41,610 --> 00:00:43,140
which can be adjusted.

9
00:00:43,410 --> 00:00:46,890
You can think of these as the same as dials on an oven.

10
00:00:46,920 --> 00:00:52,920
For example when you're learning to cook your favorite chicken dish that all that delicious sweet honey

11
00:00:52,920 --> 00:00:54,060
mustard dish.

12
00:00:54,060 --> 00:00:56,070
I'm getting hungry making these lectures.

13
00:00:56,610 --> 00:01:02,970
You notice that cooking it at 180 degrees for an hour meant it came out a little raw not ideal right.

14
00:01:02,970 --> 00:01:07,750
I want this chicken to be nice and crispy but at 200 degrees that's where we get the glory.

15
00:01:07,920 --> 00:01:09,990
That's 200 grains for an hour.

16
00:01:09,990 --> 00:01:12,000
My chicken came out just right.

17
00:01:12,000 --> 00:01:14,270
So like a little dial here Ryan.

18
00:01:14,280 --> 00:01:16,670
This is how you consume your oven.

19
00:01:16,860 --> 00:01:20,050
You can change it from 180 grains to be 200 degrees.

20
00:01:20,370 --> 00:01:25,160
If you'd lived 180 grains chicken doesn't turn out well you'd take it up 200 degrees.

21
00:01:25,170 --> 00:01:26,330
Chicken does turn out well.

22
00:01:27,430 --> 00:01:33,110
Depending on what kind of model you're using will depend on what kind of hyper properties you can chew.

23
00:01:33,130 --> 00:01:38,860
For example a random forest will allow you to adjust the number of trees in this one we've got three

24
00:01:38,860 --> 00:01:39,680
trees.

25
00:01:39,680 --> 00:01:43,240
So we've got five trees and we'll have a look at the random forest.

26
00:01:43,240 --> 00:01:45,210
This is a type of machine learning algorithm.

27
00:01:45,230 --> 00:01:52,200
We'll check this out in the future lesson and a no network will allow you to adjust the number of layers.

28
00:01:52,410 --> 00:01:58,590
Again we'll we'll look at a neural network in a future lesson just giving an example of different hyper

29
00:01:58,600 --> 00:02:01,390
parameters like the temperature of an oven.

30
00:02:01,390 --> 00:02:03,930
You can adjust on different kinds of algorithms.

31
00:02:05,090 --> 00:02:10,420
The main things for you to remember are machine learning models have hyper parameters you can adjust

32
00:02:10,940 --> 00:02:15,420
however depending on what model you're going to use the hyper parameters will be different.

33
00:02:16,320 --> 00:02:22,230
The goal of tuning hybrid parameters is to improve your model's performance and you should do model

34
00:02:22,230 --> 00:02:29,430
tuning on your training and or validation sets meaning you train a model using an initial set of hyper

35
00:02:29,430 --> 00:02:30,210
parameters.

36
00:02:30,210 --> 00:02:35,250
These might be the defaults like say for example you put your chicken ingredients in the oven you press

37
00:02:35,400 --> 00:02:41,700
chicken roast chicken on the oven and it goes for an hour on the ovens default settings and then you

38
00:02:41,700 --> 00:02:46,020
find it doesn't turn out how you want it to be you might adjust them as you go.

39
00:02:46,020 --> 00:02:48,530
This is exactly what you do with a machine learning model.

40
00:02:48,630 --> 00:02:53,440
You might try a machine learning model with an initial set of hyper parameters the default.

41
00:02:53,640 --> 00:02:56,870
And it turns out it does okay but not exactly how you want it.

42
00:02:56,880 --> 00:03:04,040
So you might try and improve it by adjusting some of the the hyper parameters up next we've got model

43
00:03:04,040 --> 00:03:04,910
comparison.

44
00:03:04,940 --> 00:03:11,150
Now this happens during your experimentation so you've trained multiple different models on the same

45
00:03:11,150 --> 00:03:12,110
dataset.

46
00:03:12,200 --> 00:03:15,520
We're going to look at how you might compare those and this is done on the test data.

47
00:03:16,270 --> 00:03:16,870
So let's do it.