1
00:00:00,490 --> 00:00:06,630
So in this we do we will learn how to find out the probability of whether a house will be sold within

2
00:00:06,630 --> 00:00:07,740
three months or not.

3
00:00:08,370 --> 00:00:10,300
Given the values of other predictors.

4
00:00:11,720 --> 00:00:16,470
So since we have already 50 model in this deal, I'm not fit very well.

5
00:00:17,190 --> 00:00:22,620
We can use this model to calculate the probabilities using the predict function.

6
00:00:24,780 --> 00:00:27,690
So we will create another variable which will take the probabilities.

7
00:00:28,020 --> 00:00:30,480
It will be called Delimit Dark from.

8
00:00:36,010 --> 00:00:37,350
Is equal to predict

9
00:00:40,090 --> 00:00:46,620
the bracket will write delimit outfit, which is the name of the evil in which we have information of

10
00:00:46,620 --> 00:00:47,760
the model comma.

11
00:00:48,900 --> 00:00:50,390
Type is equal to response.

12
00:01:00,880 --> 00:01:10,420
So you can see on the right we have this new variable delimiter probs, if you want to look at the first

13
00:01:10,420 --> 00:01:17,440
10 values of probabilities, we can IDL MDR probs from one to 10.

14
00:01:21,130 --> 00:01:24,150
So in the bottom, you can see what he first said.

15
00:01:24,410 --> 00:01:30,220
The guy's loaded probability is zero point eight five, which is very close to one, which suggests

16
00:01:30,400 --> 00:01:35,890
that the first holes will be sold within three months.

17
00:01:37,120 --> 00:01:43,340
Second value is telling us that 67 percent chance of House getting sold in the first two months.

18
00:01:45,400 --> 00:01:51,070
The third house is having only a probability of zero point three four, which means that there is only

19
00:01:51,070 --> 00:01:57,370
three percent chance of the House getting sold in the first three months and so on.

20
00:01:57,430 --> 00:01:59,280
So we have these values.

21
00:01:59,290 --> 00:02:00,020
What all leave?

22
00:02:00,070 --> 00:02:01,450
I wonder six obstetrician's.

23
00:02:04,860 --> 00:02:12,330
Now, as I told you, in the end, it's really clear that using a boundary condition, we can assign

24
00:02:12,480 --> 00:02:14,280
classes to these values.

25
00:02:14,490 --> 00:02:20,280
So if I have a boundary condition on point five, I can say that any value, any pro will devalue,

26
00:02:20,280 --> 00:02:21,630
which is greater than point five.

27
00:02:22,080 --> 00:02:23,610
I would classify it as.

28
00:02:23,940 --> 00:02:24,390
Yes.

29
00:02:24,810 --> 00:02:31,470
Meaning that it will be sold in dreamland and wherever the probability values less than point lead.

30
00:02:31,890 --> 00:02:35,220
I will say that no, it will not be sold in three months.

31
00:02:35,490 --> 00:02:40,440
So to create this area of response is containing yes and no.

32
00:02:41,190 --> 00:02:44,850
We will first create an entry called Dehlin.

33
00:02:45,000 --> 00:02:47,220
Not predict is equal to

34
00:02:50,400 --> 00:02:50,910
repeat.

35
00:02:51,060 --> 00:02:51,580
So rep.

36
00:02:52,890 --> 00:03:02,370
Basically I want to first create a array of 506 values in which only values as containing no

37
00:03:05,640 --> 00:03:06,710
comma 506.

38
00:03:07,470 --> 00:03:13,270
It's an array with 506 values and all the values are all on the right.

39
00:03:13,290 --> 00:03:20,090
You can see that we have this every now whenever the probability will lose more than point three.

40
00:03:20,970 --> 00:03:21,790
I will change this.

41
00:03:21,800 --> 00:03:22,130
Nor do.

42
00:03:22,160 --> 00:03:22,540
Yes.

43
00:03:23,130 --> 00:03:24,520
So Gillmore played.

44
00:03:28,580 --> 00:03:29,560
We good record.

45
00:03:29,670 --> 00:03:39,050
We will mention the condition that delimiter probs is greater than point five in all these locations,

46
00:03:40,730 --> 00:03:41,600
in all these locations.

47
00:03:41,630 --> 00:03:42,020
I want.

48
00:03:42,690 --> 00:03:43,130
Yes.

49
00:03:47,550 --> 00:03:50,790
So you can see the first two Aleuts I know, yes.

50
00:03:51,330 --> 00:03:51,930
The third well-used.

51
00:03:52,040 --> 00:03:52,370
No.

52
00:03:53,550 --> 00:03:54,810
Which I also discussed.

53
00:03:56,880 --> 00:04:03,780
Now that we have the predicted classes, for all the 506 observations, we can create the conclusion

54
00:04:03,780 --> 00:04:13,710
matrix confusion matrix will have predicted values on one side and the true values on other to create

55
00:04:13,710 --> 00:04:14,790
confusion matrix.

56
00:04:16,020 --> 00:04:20,040
We will use that table function and the table function.

57
00:04:20,070 --> 00:04:20,730
We have to give

58
00:04:23,690 --> 00:04:24,480
two parameters.

59
00:04:24,630 --> 00:04:30,330
The first parameter is DDL and bread values, which had the predicted values of

60
00:04:34,920 --> 00:04:36,390
whether the house will be sold or not.

61
00:04:37,080 --> 00:04:41,490
And the second is the actual value of whether the house will be sold or not.

62
00:04:41,630 --> 00:04:43,340
Which is D.F. Dollar sold.

63
00:04:47,520 --> 00:04:52,530
So this first parameter, which has the predicted values, will be coming as rules.

64
00:04:52,980 --> 00:04:57,900
And the second parameter will just be if Dollar sold will be coming at two columns.

65
00:04:58,110 --> 00:05:01,500
So let us run this in the bottom.

66
00:05:01,500 --> 00:05:02,040
You can see.

67
00:05:02,220 --> 00:05:02,540
No.

68
00:05:02,550 --> 00:05:03,180
And yes.

69
00:05:03,390 --> 00:05:06,810
Are the predicted values which we have as rules.

70
00:05:07,720 --> 00:05:13,530
Zero and one are the actual values which we have as columns.

71
00:05:14,610 --> 00:05:19,750
The first sin is telling us that one nine four one ninety seven cases.

72
00:05:21,090 --> 00:05:22,620
The actual value is zero.

73
00:05:22,800 --> 00:05:24,190
That is, the house will not be sold.

74
00:05:24,630 --> 00:05:27,180
And we also predicted that it will not be sold.

75
00:05:27,420 --> 00:05:30,480
So we were right in predicting these values.

76
00:05:31,350 --> 00:05:36,350
Similarly, these 149 cases, the actual words also one.

77
00:05:36,570 --> 00:05:37,590
And we also predicted.

78
00:05:37,590 --> 00:05:37,920
Yes.

79
00:05:38,280 --> 00:05:39,370
So that is also correct.

80
00:05:41,000 --> 00:05:44,720
But these are the two values where actual ojito.

81
00:05:44,880 --> 00:05:45,620
And we predicted.

82
00:05:45,620 --> 00:05:45,990
Yes.

83
00:05:46,700 --> 00:05:48,480
And actual odds one.

84
00:05:48,510 --> 00:05:49,530
And we really didn't know.

85
00:05:49,770 --> 00:05:53,070
So this 79 and 81, I need to edit.

86
00:05:55,650 --> 00:06:02,250
So using these few lines of code, we can get the probability of that case belonging to each class,

87
00:06:03,390 --> 00:06:05,210
then using a ball degradation.

88
00:06:05,370 --> 00:06:15,420
We can assign a class to each observation using those predicted values of classes and the actual value

89
00:06:15,420 --> 00:06:15,990
of classes.

90
00:06:16,140 --> 00:06:22,170
We can draw this conclusion matrix to see how well our model is predicting the response where he would.