1
00:00:00,290 --> 00:00:07,140
All in all our previous sessions, we have done lots of analysis, lots of cleaning, lots of techniques

2
00:00:07,350 --> 00:00:12,970
for future importance, feature encoding and much, much more so in this session.

3
00:00:13,260 --> 00:00:16,470
This is the session that you all guys are waiting for.

4
00:00:16,720 --> 00:00:23,340
And in this session, we are going to apply our machine learning algorithm on data because up to a greater

5
00:00:23,340 --> 00:00:28,900
extent, our data is ready to apply your machine learning algorithm on your data.

6
00:00:29,220 --> 00:00:31,100
So let's open your assignment.

7
00:00:31,110 --> 00:00:32,010
What you have to do.

8
00:00:32,370 --> 00:00:37,200
The very first statement is you have to apply an algorithm after it.

9
00:00:37,200 --> 00:00:39,840
You have to cross validate your model as well.

10
00:00:40,380 --> 00:00:47,070
So for this, what we have to do very first, you have to split your data in the form of training as

11
00:00:47,070 --> 00:00:52,100
well as your testing, because once you have trained data, you can train your model.

12
00:00:52,440 --> 00:00:57,480
Once you have test data and you can check what exactly is the accuracy of your model.

13
00:00:57,900 --> 00:01:04,170
So for this, I'm just going to say from my Escalon module, I have something which is exactly my model

14
00:01:04,170 --> 00:01:05,160
and of course, selection.

15
00:01:05,400 --> 00:01:11,430
And from this I have to import something known as train on a score test and a score split just to execute

16
00:01:11,430 --> 00:01:11,520
it.

17
00:01:11,550 --> 00:01:15,240
Now, what we have to do is you have to simply call this class.

18
00:01:15,240 --> 00:01:21,540
And in this, if I am going to pass shift plus tab, you will check all these parameters over here,

19
00:01:21,540 --> 00:01:28,380
Taci train side Sefl, Stratify splitting and all these different, different things.

20
00:01:28,380 --> 00:01:34,920
And what will it and it will me this exchange test y train and whitelist.

21
00:01:35,430 --> 00:01:42,780
So here I'm just going to say I have to pass my X and Y and after it I have to pass my test side.

22
00:01:42,780 --> 00:01:47,890
So I'm going to say test on the cost side, let's say nothing or zero point two five.

23
00:01:47,910 --> 00:01:53,190
It means you're twenty five percent of data will be considered as a test data and the rest will be you're

24
00:01:53,550 --> 00:01:54,220
carrying it out.

25
00:01:54,230 --> 00:01:57,930
So I'm just going to say it will be done with all these parameters.

26
00:01:57,930 --> 00:02:05,820
So I'm just going to execute this cell and let me pass one more parameter, which is exactly random

27
00:02:05,820 --> 00:02:12,290
underscore estate, which is exactly, I'm going to say random on a score instead and just a standard

28
00:02:12,300 --> 00:02:14,380
zero and just execute this.

29
00:02:14,520 --> 00:02:21,190
Now, what we have to do, we have to simply apply our machine learning algorithm on our data.

30
00:02:21,570 --> 00:02:28,200
So let's say we are just going to apply logistic regression because this is exactly of a classification

31
00:02:28,200 --> 00:02:28,940
algorithm.

32
00:02:29,250 --> 00:02:33,080
So here we have to apply logistic regression on our data.

33
00:02:33,090 --> 00:02:37,200
And if you guys are not that much of it about what exactly the velocity regression.

34
00:02:37,410 --> 00:02:42,540
So I have explained the mathematics intuition behind this logistic regression.

35
00:02:42,870 --> 00:02:50,100
Just check out my previous videos and you will be able to find what exactly is and mathematics, how

36
00:02:50,100 --> 00:02:53,940
this logistic regression exactly works in your machine learning.

37
00:02:54,150 --> 00:02:57,900
So very first, I have to aim for something which is exactly my own site.

38
00:02:57,910 --> 00:03:01,440
Gatland got linear underscore model.

39
00:03:01,440 --> 00:03:04,550
I have to import my lines to progression.

40
00:03:04,560 --> 00:03:07,140
So I'm just going to say just execute the cell.

41
00:03:07,210 --> 00:03:10,430
Now, what we have to do, we have to initialize this logistic regression.

42
00:03:10,440 --> 00:03:17,700
So I'm going to say logistic regression just in size it let's say here you have some custom parameters

43
00:03:17,700 --> 00:03:17,910
here.

44
00:03:17,910 --> 00:03:21,520
You can simply say you have dozens of parameters over here.

45
00:03:21,810 --> 00:03:28,080
So let me iStore its object as, let's say, Lavarack, just execute the cell.

46
00:03:28,120 --> 00:03:33,120
Now, what we have to do using this object, I'm just going to call my fit over there and what I have

47
00:03:33,120 --> 00:03:43,140
to fit I have to fit my training data, which is exactly X on this score and basically my Y on the screen

48
00:03:43,140 --> 00:03:45,330
because I have a screen model.

49
00:03:45,330 --> 00:03:45,870
That's it.

50
00:03:46,080 --> 00:03:47,850
And just execute the set.

51
00:03:47,850 --> 00:03:53,040
It will take a while and it will return as my amazing logistic regression model.

52
00:03:53,040 --> 00:03:55,940
You will see all these custom parameters over here.

53
00:03:55,950 --> 00:03:58,870
What is the parameter for the class.

54
00:03:58,890 --> 00:04:01,850
Wait, what is it all your different different parameters?

55
00:04:02,100 --> 00:04:06,800
What exactly is your regularization technique, which is exactly all these things.

56
00:04:07,170 --> 00:04:11,750
Now what you have to do, you have to simply do prediction on your testing data.

57
00:04:11,760 --> 00:04:16,860
So I'm just going to say LaGreca not predict what I have to predict.

58
00:04:16,860 --> 00:04:22,380
I have to predict on my X on a test after doing prediction, let's say I have restored my prediction

59
00:04:22,380 --> 00:04:25,710
X, I'm going to say it is nothing but my yna school grad.

60
00:04:25,950 --> 00:04:27,180
It just executed.

61
00:04:27,180 --> 00:04:33,270
And if I have to paint this so just execute this, sell all this stuff gets executed over it, you will

62
00:04:33,270 --> 00:04:37,650
see all the prediction in the form of add it over here to wherever it is one.

63
00:04:37,890 --> 00:04:43,560
It means your book is going to cancel and wherever it is zero, it means your booking is not going to

64
00:04:43,560 --> 00:04:43,920
cancel.

65
00:04:44,010 --> 00:04:48,270
That's why it is exactly what classification use is now.

66
00:04:48,270 --> 00:04:55,490
What we have to do, we just need what exactly is a confusing matrix of my logistic regression model.

67
00:04:55,830 --> 00:04:59,940
So for this, I'm just going to say from this side, Gittleman, I.

68
00:05:00,140 --> 00:05:07,190
Playing board, something which is exactly my metrics, and from these metrics, I have to import something

69
00:05:07,190 --> 00:05:08,760
which is my confusion metrics.

70
00:05:09,050 --> 00:05:15,470
Now what we have to do from this confusion call metrics, I have to pass my ad, which is exactly my

71
00:05:15,830 --> 00:05:17,150
lie and test.

72
00:05:17,480 --> 00:05:22,280
The second is exactly my why on this score, Brad just executed.

73
00:05:22,280 --> 00:05:24,220
And this is your amazing metrics.

74
00:05:24,230 --> 00:05:30,740
You will see these are exactly the correct prediction as well as these are exactly correct predictions

75
00:05:30,740 --> 00:05:34,100
and these are your not correct predictions.

76
00:05:34,340 --> 00:05:35,420
So let's see.

77
00:05:35,420 --> 00:05:41,120
I just need to hear what exactly the accuracy of my model so far is.

78
00:05:41,120 --> 00:05:41,900
What I'm going to do.

79
00:05:41,900 --> 00:05:51,320
I'm just going to say from this cyclone dot matrix, I have to import something known as which is exactly

80
00:05:51,320 --> 00:05:52,400
accuracy score.

81
00:05:52,760 --> 00:05:56,090
Now, what I have to do, I'm just going to say accuracy score.

82
00:05:56,360 --> 00:05:58,790
And in this accuracy score, I have to pass.

83
00:05:58,820 --> 00:06:00,470
Why test after it?

84
00:06:00,470 --> 00:06:04,510
I have to pass a wide spread and just execute the cell.

85
00:06:04,520 --> 00:06:11,380
You will see your Lustick regression model has somewhere close to seventy two percent accuracy.

86
00:06:11,420 --> 00:06:17,230
So if you guys are thinking this is the exact accuracy of my last two regression model.

87
00:06:17,570 --> 00:06:19,420
So no, you all are wrong.

88
00:06:19,430 --> 00:06:23,720
This is not that accuracy which is related to velocity regression model.

89
00:06:23,720 --> 00:06:31,100
It means you have to come across with your right accuracy so how you can achieve that exact accuracy.

90
00:06:31,130 --> 00:06:34,910
So in such scenarios, you have to cross validate your model.

91
00:06:35,240 --> 00:06:36,470
So let me show you a thing.

92
00:06:36,470 --> 00:06:42,050
If I'm going to let's say if I'm going to change this random sandwiches and its value as, let's say,

93
00:06:42,050 --> 00:06:42,560
42.

94
00:06:42,980 --> 00:06:50,240
So if I'm going to execute again, execute again and let me just execute again, let me show you a thing.

95
00:06:50,810 --> 00:06:57,980
And if again, I'm going to execute it and I'll get executed and let me show you what exactly is accuracy

96
00:06:57,980 --> 00:07:01,130
after playing with my randomness Christian parameter.

97
00:07:01,340 --> 00:07:03,940
And again, I'm going to execute the search.

98
00:07:03,960 --> 00:07:10,630
Now you will see your accuracy from seventy one gets jump to seventy five.

99
00:07:11,000 --> 00:07:12,730
So you will thinking yeah.

100
00:07:12,770 --> 00:07:14,370
You automatical going over here.

101
00:07:14,720 --> 00:07:17,720
So that's the scenario where you have to cross my little model.

102
00:07:18,140 --> 00:07:25,040
So whenever you will get your accuracy without consolidating your model, you have to always say I have

103
00:07:25,040 --> 00:07:27,830
achieved accuracy within this, this range.

104
00:07:28,820 --> 00:07:35,270
It means now to get the exact accuracy you have to cross, validate your model for this.

105
00:07:35,270 --> 00:07:41,480
I'm just going to say from this side, Keadilan, I have something known as model selection.

106
00:07:41,870 --> 00:07:45,110
So I'm going to say model unaskable selection.

107
00:07:45,110 --> 00:07:52,250
And from this I have to import something known as cross on, as Corvalan is called, just execute the

108
00:07:52,250 --> 00:07:52,750
cell.

109
00:07:52,790 --> 00:07:58,070
Now, what we have to do that is first I have to initialize it and just shift crosstab.

110
00:07:58,070 --> 00:07:58,580
You will see.

111
00:07:59,150 --> 00:08:00,700
What exactly is your estimate?

112
00:08:00,920 --> 00:08:01,680
What is the X?

113
00:08:01,730 --> 00:08:05,340
What is it Y and what is your CV parameter?

114
00:08:05,360 --> 00:08:12,050
So if you guys doesn't know what exactly this cross validation, just check out my previous videos.

115
00:08:12,050 --> 00:08:14,010
I have explained it and everything.

116
00:08:14,010 --> 00:08:14,580
Go there.

117
00:08:14,960 --> 00:08:18,630
So here, if you will pass through Gustav, you will get all these parameters over here.

118
00:08:18,650 --> 00:08:21,820
Now you have to set all these stuff to my estimate.

119
00:08:21,930 --> 00:08:30,250
What exactly Lorac and my X was X, my life was like, let's say I have to assign my cross-validation.

120
00:08:30,280 --> 00:08:31,700
It goes to ten.

121
00:08:32,000 --> 00:08:37,840
And if let's say whatever it will on me, let's say I'm just going to execute the set.

122
00:08:37,880 --> 00:08:46,070
Now, you will see over here with respect to X, Y and the C.V parameter, you have this, this is called

123
00:08:46,290 --> 00:08:47,840
selectees told somewhere else.

124
00:08:47,840 --> 00:08:52,640
Let's say I'm just going to store it in score and let me just execute it again.

125
00:08:52,940 --> 00:09:01,340
Now, what we have to do to achieve your exact score, you guys can just call me an order that simple.

126
00:09:01,340 --> 00:09:03,330
You have to just execute the cell.

127
00:09:03,350 --> 00:09:10,240
And now you will see over here it has accuracy of somewhere close to 70 percent.

128
00:09:10,250 --> 00:09:14,240
It means you're 70 percent predictions are going to.

129
00:09:14,240 --> 00:09:14,880
Correct.

130
00:09:14,900 --> 00:09:17,780
So that's all about the session of the session.

131
00:09:17,780 --> 00:09:24,580
Very much how exactly I have implemented my machine learning algorithm on data, then how I have crossed

132
00:09:24,640 --> 00:09:32,240
validate my model as well, because this is your most important aspect of the machine learning approach.

133
00:09:32,720 --> 00:09:36,020
So I hope you love this session and this session will be very helpful for you.

134
00:09:36,440 --> 00:09:37,080
Thank you.

135
00:09:37,130 --> 00:09:38,070
Have a nice day.

136
00:09:38,090 --> 00:09:38,960
Keep learning.

137
00:09:38,960 --> 00:09:39,830
Keep growing.

138
00:09:40,040 --> 00:09:40,820
Keep practicing.