1
00:00:01,410 --> 00:00:06,420
Now let's learn how to create a logistic regression model in Python.

2
00:00:07,500 --> 00:00:11,340
So in Python, there are two ways to create logistic regression model.

3
00:00:11,580 --> 00:00:13,760
First, using Escalon libraries.

4
00:00:14,190 --> 00:00:17,870
And second method is using state smarter library.

5
00:00:19,590 --> 00:00:25,550
There are a lot of documentation and implementation techniques available for a skill.

6
00:00:26,160 --> 00:00:28,740
That's why it is more preferred by professionals.

7
00:00:29,450 --> 00:00:32,880
Whereas a site model is the new library.

8
00:00:32,970 --> 00:00:38,340
And there are hardly any documentation and there are some bugs as well.

9
00:00:39,120 --> 00:00:45,660
But search model provides all the statistical information that is needed from all logistic regression

10
00:00:45,660 --> 00:00:45,960
model.

11
00:00:46,770 --> 00:00:50,340
So we will create this model using both of these two libraries.

12
00:00:54,210 --> 00:00:59,140
But before that, we need to create our X and Y variables.

13
00:01:01,990 --> 00:01:05,410
This time, we are only selecting one independent variable.

14
00:01:06,810 --> 00:01:12,600
That is price and our vye dependent variable is the Soviet variable.

15
00:01:14,020 --> 00:01:15,640
So first.

16
00:01:17,360 --> 00:01:24,080
We will clear our equitable will, right x equity to be if.

17
00:01:28,500 --> 00:01:39,210
And then to square record, we will right price in Eskil on your ex or independent variable should be

18
00:01:39,210 --> 00:01:40,470
a two dimensional.

19
00:01:40,920 --> 00:01:43,530
That's why we are using two squared records.

20
00:01:47,400 --> 00:01:49,970
And the word via variable is sold.

21
00:01:50,400 --> 00:01:51,840
So we'll write like way to.

22
00:02:00,750 --> 00:02:04,130
Sorry and single this quarter.

23
00:02:04,550 --> 00:02:05,540
We will rate soared.

24
00:02:08,470 --> 00:02:09,490
Now, let's.

25
00:02:10,720 --> 00:02:16,090
Check the sample of these two data frames, right, x dopehead.

26
00:02:26,030 --> 00:02:33,050
We have only one variable, which is players in the league, we will check, assemble or fly.

27
00:02:39,490 --> 00:02:48,490
So first, we will create our model using a scale, then we have to import logistic regression from

28
00:02:48,490 --> 00:02:49,090
a Skillern.

29
00:02:49,510 --> 00:02:52,090
So we will write from Escalon totally Nantel Mardon.

30
00:03:00,980 --> 00:03:02,700
In both logistic regression.

31
00:03:09,730 --> 00:03:12,820
You don't have to remember all this Gord's.

32
00:03:15,180 --> 00:03:21,930
You can visit Eskil Skillern documentation or you can copy this skorts from this notebook, which is

33
00:03:21,990 --> 00:03:24,120
a test in the resource section of this video.

34
00:03:26,200 --> 00:03:30,010
For Escalon, we followed three steps.

35
00:03:30,400 --> 00:03:32,710
First, we create our classification object.

36
00:03:33,400 --> 00:03:37,300
Then we find that object using our X and Y variable.

37
00:03:37,990 --> 00:03:41,770
And the third step is to predict our Y variable.

38
00:03:41,980 --> 00:03:44,380
Using that drained classification model.

39
00:03:46,840 --> 00:03:50,140
So first we will create our object.

40
00:03:50,570 --> 00:03:58,900
CLV underscore a lattice CLV is for classification and allow this sense for logistic regression.

41
00:03:59,560 --> 00:04:02,710
And we are using logistic regression with single variable.

42
00:04:02,770 --> 00:04:04,930
That's why we are putting in this word.

43
00:04:04,930 --> 00:04:08,810
C11 underscore a is a variable name.

44
00:04:10,300 --> 00:04:14,200
And we won this object to take logistic regression.

45
00:04:30,300 --> 00:04:34,070
Now we will fit our X and Y with even Buffett.

46
00:04:34,170 --> 00:04:44,130
We just have the right outfit, so we'll write CLV, underscore LARUS, not fit, and then record will

47
00:04:44,280 --> 00:04:51,060
first mention our dependent variable, which is X and then our dependent variable, which is why.

48
00:04:56,980 --> 00:05:03,870
We'll run this, we have grain, though, what a model to check the coefficient and intercept of a model

49
00:05:04,000 --> 00:05:06,280
that is B does little and B Devon.

50
00:05:07,750 --> 00:05:11,230
I hope you remember we does it when we govern from a word to reflect reelected.

51
00:05:12,550 --> 00:05:16,450
We'll see value of B does it when we do when we can.

52
00:05:17,200 --> 00:05:17,440
Right.

53
00:05:18,440 --> 00:05:20,060
CLV underscore Allardice.

54
00:05:24,950 --> 00:05:33,380
Not if underscore or if it stands for coefficient, and this is the attribute of over modern.

55
00:05:35,120 --> 00:05:37,170
We on this, we are getting this value.

56
00:05:37,240 --> 00:05:44,910
This is a value of between to see the value of B to zero, relates VLF.

57
00:05:45,040 --> 00:05:49,950
On that score, a lot is not intercept underscore.

58
00:05:54,350 --> 00:06:02,440
Again, this school, if and intercepts are the attributes of forward logistic regression model.

59
00:06:05,370 --> 00:06:07,210
There's the value of zero.

60
00:06:10,830 --> 00:06:19,250
So using this bit does it when we touch one, you can create a logistic equation that we discuss in

61
00:06:19,260 --> 00:06:22,200
our theory lecture and you can use this.

62
00:06:22,760 --> 00:06:27,280
What does it when we don't need to predict the value of Y from here when news of X?

63
00:06:28,620 --> 00:06:29,910
This is the first method.

64
00:06:30,880 --> 00:06:32,680
Of creating logistic regression model.

65
00:06:33,010 --> 00:06:38,710
Now let's move on to the second method, which is using a service model to create logistic regression.

66
00:06:40,690 --> 00:06:51,730
Now, one drawback of using a set model is that by default, set model do not use a consent form, which

67
00:06:51,730 --> 00:06:55,540
means that you are does zero will be zero by default.

68
00:06:57,120 --> 00:07:04,410
So to add concern, Tom, and do a lot more than we first need to do, concern them in a word, dependent

69
00:07:04,410 --> 00:07:05,100
variables.

70
00:07:07,160 --> 00:07:11,450
To do that, we will first import, said Smardon Lord Apia.

71
00:07:19,140 --> 00:07:20,160
As Esson.

72
00:07:25,220 --> 00:07:25,780
We'll run it.

73
00:07:32,850 --> 00:07:38,380
No, we'll create another door frame for our dependent variable, which is X under Scott Korsnes.

74
00:07:39,140 --> 00:07:39,620
This.

75
00:07:40,580 --> 00:07:43,850
Contain one single, call them as a constant.

76
00:07:47,130 --> 00:07:47,500
All right.

77
00:07:47,530 --> 00:07:51,910
Xander Scott wants equally to ascendent at constant.

78
00:08:04,560 --> 00:08:06,030
Again, you don't have to.

79
00:08:06,940 --> 00:08:14,590
Remember all this codes, you can copy this code and follow this code for your future projects.

80
00:08:16,760 --> 00:08:23,930
Let's again, we will see a sample of four X concent will right x kohn's da tagged.

81
00:08:33,650 --> 00:08:37,940
You can see there is one consent I did in two or X.

82
00:08:41,460 --> 00:08:47,220
Now to import the logistic regression classifier from a Setswana right?

83
00:08:47,640 --> 00:08:48,240
Import.

84
00:08:51,790 --> 00:08:55,930
Then sets modern, discreet, not discreet, modern.

85
00:09:04,640 --> 00:09:08,130
And we will import this as s.m.

86
00:09:10,950 --> 00:09:11,830
We'll run this.

87
00:09:14,410 --> 00:09:23,260
Now we need to screen our X and Y evitable and do a model, so we'll write logic lodged as our variable

88
00:09:23,260 --> 00:09:23,560
name.

89
00:09:26,190 --> 00:09:27,680
s.M, not logic.

90
00:09:35,870 --> 00:09:38,470
Here first, we need to mention our white variable.

91
00:09:38,960 --> 00:09:45,060
We will write why and then our X is X underscore Korsnes.

92
00:09:46,790 --> 00:09:50,390
Remember, we are adding a constant thumb and door X variable.

93
00:09:53,660 --> 00:09:55,420
And then we need to fit.

94
00:10:00,320 --> 00:10:03,230
Fit on this auto model, is that Audi?

95
00:10:05,580 --> 00:10:07,260
To see the somebody off a little more than.

96
00:10:09,070 --> 00:10:10,970
We'll wait logic, but somebody.

97
00:10:26,620 --> 00:10:31,480
Here you can see all the information about the logistic regression model.

98
00:10:32,290 --> 00:10:34,620
We are only getting this using a search model.

99
00:10:34,900 --> 00:10:37,480
We were not getting this using a Skillern.

100
00:10:40,950 --> 00:10:47,180
Here you can see at the bottom, we have Budos forces for consent.

101
00:10:47,310 --> 00:10:48,930
And second one is for Kreiss.

102
00:10:51,100 --> 00:10:59,500
Then in the first column, we go, if this are the coefficient, the coefficient of consent is also

103
00:10:59,500 --> 00:11:05,170
known as does zero and this the coefficient of price will be a lot to be done.

104
00:11:06,850 --> 00:11:11,010
We have a standard at our values corresponding to this coefficient.

105
00:11:11,950 --> 00:11:14,650
And then we have the Z value for this confusion.

106
00:11:18,570 --> 00:11:21,790
In the fourth column, we have the P-value.

107
00:11:24,160 --> 00:11:29,650
We will discuss about the standard error ZEWAIL You and the P-value.

108
00:11:29,950 --> 00:11:31,420
And our next theory lecture.

109
00:11:36,390 --> 00:11:41,790
On the top of this fable, you can see or deliver dependent variable.

110
00:11:42,130 --> 00:11:43,160
What is the model?

111
00:11:43,560 --> 00:11:51,060
And my third, remember, my third is maximum likelihood estimate that that's what it is.

112
00:11:51,200 --> 00:11:51,460
M.

113
00:11:51,530 --> 00:11:51,740
L.

114
00:11:51,930 --> 00:11:52,290
E.

115
00:11:53,610 --> 00:11:55,620
Then we have no observations.

116
00:11:55,740 --> 00:12:00,270
The residuals and other details of a logistic regression model.

117
00:12:02,320 --> 00:12:05,320
So remember, we are getting this information online.

118
00:12:05,410 --> 00:12:06,280
It start Smardon.

119
00:12:08,380 --> 00:12:12,660
And start small, we can see all this information about our model.

120
00:12:13,360 --> 00:12:20,440
But professionals use a Skillern for most of their work because of the high quality of documentation

121
00:12:20,830 --> 00:12:24,100
and the additional features that are present in Escalon.
