1
00:00:01,530 --> 00:00:05,380
So now let us see how to create a random forest model in our.

2
00:00:07,470 --> 00:00:11,610
Again, to create random forest model, we need to use the package.

3
00:00:11,760 --> 00:00:12,570
Random forest.

4
00:00:13,250 --> 00:00:16,470
This is the same package we use to do bagging and Jawdat.

5
00:00:16,500 --> 00:00:19,380
This package is installed and it is active.

6
00:00:20,500 --> 00:00:27,160
So since this package started for me, I'll just run this Libatique a one to make sure that it is active

7
00:00:27,160 --> 00:00:27,580
for me.

8
00:00:28,810 --> 00:00:34,510
Next step is to use the land random forest function to create the model random for is the name of my

9
00:00:34,510 --> 00:00:38,230
variable, which will contain the information from this random forest function.

10
00:00:39,650 --> 00:00:41,270
The first barometer is the formula.

11
00:00:41,870 --> 00:00:48,560
This is the same collection as my dependent variable, all of that on my independent ready world data

12
00:00:48,590 --> 00:00:49,700
is brain data.

13
00:00:50,810 --> 00:00:58,400
And three is the barometer, which specifies the maximum number of threes that this model should create.

14
00:01:00,660 --> 00:01:09,780
If you remember in bagging, we used a barometer called M3 and we specified that Emtriva value should

15
00:01:09,780 --> 00:01:12,840
be equal to all the other independent variables.

16
00:01:12,960 --> 00:01:14,730
So we equated it to 17.

17
00:01:15,990 --> 00:01:23,130
In this scenario, since Random Forest considers less number of rebuild, that is a subset of the total

18
00:01:23,140 --> 00:01:24,300
predicted space.

19
00:01:25,540 --> 00:01:29,620
Therefore, either we specify Emtriva less than 17.

20
00:01:30,740 --> 00:01:37,760
So maybe we can say M3 is equal to five or embrace equal to six, something like that, or we do not

21
00:01:37,760 --> 00:01:38,660
specify anything.

22
00:01:38,780 --> 00:01:40,790
And are will use a default value.

23
00:01:41,320 --> 00:01:47,040
So by default, since this is a regression model, I will use PMA three.

24
00:01:47,750 --> 00:01:49,010
Since we have 17.

25
00:01:50,820 --> 00:01:51,720
Predictor variables.

26
00:01:51,770 --> 00:01:55,080
B by three will come out to be five point six six.

27
00:01:56,530 --> 00:01:57,520
Which is nearly six.

28
00:01:57,670 --> 00:02:01,060
So probably ordering take em crazy equal to six.

29
00:02:03,010 --> 00:02:10,090
If you want to use other value, you can just write a comma and the best way M3 is equal to some of

30
00:02:10,090 --> 00:02:11,470
that value which you want to use.

31
00:02:12,070 --> 00:02:13,030
So I love this model.

32
00:02:17,470 --> 00:02:22,070
And random forward variable has the information of the random forest model.

33
00:02:24,450 --> 00:02:31,350
Now, using the predicate function I predicted devalues and using this formula for MASC.

34
00:02:31,620 --> 00:02:34,140
I calculated the MSE for Random Forest.

35
00:02:35,820 --> 00:02:43,160
Now you can see that a messy for random forest is coming out to be 41 million, nearly four bagging,

36
00:02:43,170 --> 00:02:45,420
it was 54, 55 million.

37
00:02:45,780 --> 00:02:49,340
So you can clearly see an improvement in the model.

38
00:02:50,880 --> 00:02:57,000
There is nearly a 20 percent improvement in terms of masc of the predicted values.

39
00:02:58,980 --> 00:03:06,750
So just by changing the value of M3 in this whole process, that is instead of using Embry's equal to

40
00:03:06,750 --> 00:03:07,470
17.

41
00:03:09,010 --> 00:03:11,530
We are using embryos equal to nearly six.

42
00:03:12,040 --> 00:03:17,210
I am getting a huge improvement in prediction, accuracy of my model.

43
00:03:18,520 --> 00:03:24,140
And by the way, I hope since I'm doing bagging and rainforest's only for recreation, I hope that you're

44
00:03:24,160 --> 00:03:26,530
doing it for classification on your own.

45
00:03:27,220 --> 00:03:29,800
You know how to run that classification model.

46
00:03:30,190 --> 00:03:33,580
Only thing you need to add is another parameter here, which is method.

47
00:03:34,180 --> 00:03:38,350
So that works for bagging and random forest as well.

48
00:03:38,590 --> 00:03:44,140
So my suggestion is that you should go ahead and do this for classification also.

49
00:03:45,210 --> 00:03:48,200
And see the improvement in classification prediction accuracy.

50
00:03:49,140 --> 00:03:55,200
So do this work classification, make their confusion metrics and see the improvement in prediction,

51
00:03:55,200 --> 00:03:57,510
accuracy, using, bagging and random forest.
