1
00:00:01,630 --> 00:00:03,760
Hello, everyone, and last lecture.

2
00:00:03,880 --> 00:00:07,570
We created a simple linear regression model in Python.

3
00:00:08,080 --> 00:00:12,760
We use a site, Model API and a Skillern to create that model.

4
00:00:13,660 --> 00:00:17,190
Now let's start building them or deepen the integration model.

5
00:00:17,430 --> 00:00:18,100
And by then.

6
00:00:19,600 --> 00:00:26,990
First, we will build our model using Start Smardon, and then we will again be using Eskinder.

7
00:00:30,020 --> 00:00:33,810
I have already imported my estate's model and Escalon liabilities.

8
00:00:34,640 --> 00:00:36,140
So if you are planning this.

9
00:00:37,070 --> 00:00:39,650
Don't forget to import the required libraries.

10
00:00:41,110 --> 00:00:48,160
Now, similar to last thing, we need to first create X and Y variables for over more than.

11
00:00:49,670 --> 00:00:51,920
So let's first clear dollar, swift even.

12
00:00:53,440 --> 00:00:59,530
Last time, we only use rule number as our x ray, but this time we need all the variables except the

13
00:00:59,530 --> 00:01:00,230
price variable.

14
00:01:00,520 --> 00:01:07,110
Since Bizer table is our dependent variable, we need all the variables except price variable from a

15
00:01:07,110 --> 00:01:10,510
lot B if as our X would even write X.

16
00:01:11,460 --> 00:01:13,000
Underscore my date.

17
00:01:13,720 --> 00:01:16,930
This is our variable name for dependent variables.

18
00:01:17,610 --> 00:01:21,180
We will write B if not go.

19
00:01:23,410 --> 00:01:26,710
And then record will right price.

20
00:01:29,730 --> 00:01:32,940
And after a coma, we'll write Exis Equal the one.

21
00:01:36,490 --> 00:01:39,180
Let me explain you discovery first.

22
00:01:39,370 --> 00:01:42,090
We are creating a variable X underscore morality.

23
00:01:43,240 --> 00:01:50,170
Then we want all the variables of our beef data frame in this variable except the price variable.

24
00:01:50,770 --> 00:01:51,900
So we are using graph.

25
00:01:51,940 --> 00:01:54,520
Come on, we'll grab the guys column name.

26
00:01:55,880 --> 00:02:00,950
And since we are dropping a call them, we need to mention access equally to one.

27
00:02:02,030 --> 00:02:07,820
If we were dropping any rules, so, for example, if I wanted to drop my first school.

28
00:02:09,370 --> 00:02:13,750
I would use Grop and then in record I will write one.

29
00:02:14,610 --> 00:02:16,780
And in excess I would mention zero.

30
00:02:17,430 --> 00:02:23,410
X is zero is for dropping rose and X is equal to one is for dropping columns.

31
00:02:26,460 --> 00:02:27,890
So and this is statement.

32
00:02:28,560 --> 00:02:34,430
We are just dropping price and saving out all the variables and Door X multivariable.

33
00:02:38,250 --> 00:02:40,740
If we take a head of this X, my D.

34
00:02:42,010 --> 00:02:42,630
DUFFING.

35
00:02:49,190 --> 00:02:52,040
This is a sample of what it's my duty to frame.

36
00:02:52,760 --> 00:02:56,070
And you can see price sweaty when they scrub and desert.

37
00:02:57,590 --> 00:03:00,410
We only have our independent variable in this is a.

38
00:03:04,470 --> 00:03:07,980
Now, let's clear Tuller dependent variable name.

39
00:03:08,370 --> 00:03:09,870
Why are indiscriminately.

40
00:03:14,510 --> 00:03:19,110
And since our prize is over, the Bindoon video, but will right be off?

41
00:03:20,060 --> 00:03:22,130
And in this good record with mentioned price.

42
00:03:25,050 --> 00:03:27,330
We can also take a sample of this.

43
00:03:37,420 --> 00:03:42,520
You can see now we have created a what dependent and independent variables.

44
00:03:44,580 --> 00:03:53,180
If you remember and Isaak's Smardon, we have to act concerned, deliver data by default, what be done.

45
00:03:53,400 --> 00:03:55,530
Which is consent is zero.

46
00:03:56,000 --> 00:03:56,820
And they said smart.

47
00:03:57,750 --> 00:04:01,860
Let's add a constant to the word dependent variable x underscore Moradi.

48
00:04:03,570 --> 00:04:05,150
We'll create another debate.

49
00:04:05,310 --> 00:04:08,160
But X underscore my date under this good Korn's.

50
00:04:15,360 --> 00:04:18,570
All right, write Essene dot edcon spent.

51
00:04:23,180 --> 00:04:27,110
And then record will mention our x ray was, which is x mightly.

52
00:04:33,470 --> 00:04:37,010
If we take a sample of this, X might be Korn's.

53
00:04:43,560 --> 00:04:47,510
You can see, Ali, at our first columbus' crime rate.

54
00:04:48,030 --> 00:04:52,370
Now, our first column is this concept and this is taking Relu one.

55
00:04:53,070 --> 00:04:55,950
So this is a kind of proxy for what would be dunnart.

56
00:04:59,550 --> 00:05:00,960
Now to for Total Waman.

57
00:05:02,510 --> 00:05:07,460
Will create an alarm object, will write alarm, and that's called Moradi.

58
00:05:08,930 --> 00:05:10,610
Again, this is our variable name.

59
00:05:11,270 --> 00:05:17,750
You can say a word, object name, and we want to fit ordinary Lisa Square.

60
00:05:17,920 --> 00:05:25,510
So we'll write Esson Dot or as or unless a sense for ordinary Lisa Square, it is a method to make a

61
00:05:25,600 --> 00:05:26,210
patient model.

62
00:05:27,420 --> 00:05:29,300
We'll mention why underscore might be.

63
00:05:32,360 --> 00:05:36,530
And our X is X underscore may be underscored Conn's.

64
00:05:39,000 --> 00:05:41,310
There's the one in which we edit consent.

65
00:05:43,640 --> 00:05:44,790
Then we like big.

66
00:05:48,380 --> 00:05:49,160
Good on this.

67
00:05:51,270 --> 00:05:57,320
We have today a lot more that in this Elmendorf's got a mighty good view, this Smardon, right?

68
00:05:57,780 --> 00:06:00,060
I mean, let's go to my day then.

69
00:06:00,200 --> 00:06:00,430
Right.

70
00:06:00,500 --> 00:06:01,320
Got somebody.

71
00:06:08,400 --> 00:06:10,680
This is the result of a lot more than.

72
00:06:12,660 --> 00:06:16,720
You can see the artist square value or model is zero points.

73
00:06:16,770 --> 00:06:17,460
I went to one.

74
00:06:18,620 --> 00:06:22,370
And it just said artists good value is zero points, I wouldn't want to.

75
00:06:24,380 --> 00:06:25,230
Which is pretty good.

76
00:06:27,580 --> 00:06:33,400
Then about probability of F statistics is very low.

77
00:06:33,940 --> 00:06:38,000
It's in the range of eight to the bottom minus 125.

78
00:06:39,280 --> 00:06:47,110
So we can say with confidence that our independent variables have some impact on our dependent variable.

79
00:06:50,540 --> 00:07:01,370
The total number of observation network data are 506 and degrees of freedom is 490, since the formula

80
00:07:01,370 --> 00:07:04,220
for calculating the degrees of freedom is an.

81
00:07:05,170 --> 00:07:08,570
Minus B, minus one zero.

82
00:07:08,740 --> 00:07:11,390
And this five zero six would be.

83
00:07:11,710 --> 00:07:16,630
That is number of variables is 15 since we have been independent variables.

84
00:07:17,740 --> 00:07:23,850
And then minus one, that is five hundred and six, minus 16, which will give us four 90.

85
00:07:23,930 --> 00:07:26,020
So what degree of freedom is for 90?

86
00:07:28,150 --> 00:07:31,630
Now, let's look at individual independent variables.

87
00:07:33,790 --> 00:07:39,910
You can see all of what independent variables are listed on the left.

88
00:07:41,410 --> 00:07:43,830
The next column is, of course, fishing.

89
00:07:44,770 --> 00:07:50,500
These are the values of be done or be done with two, three and so on.

90
00:07:51,200 --> 00:07:52,360
Bita fifteen.

91
00:07:55,430 --> 00:07:57,200
Next is the standard edit.

92
00:07:58,670 --> 00:07:59,830
After that, we have a D.

93
00:07:59,990 --> 00:08:03,590
Statistics associated with this variables.

94
00:08:05,280 --> 00:08:07,440
Then we have a P-value of these variables.

95
00:08:08,160 --> 00:08:08,910
So remember.

96
00:08:09,970 --> 00:08:15,210
The lower the up you add, Lou, the more significant the variable is in determining way.

97
00:08:19,800 --> 00:08:21,970
So if you see the lowest.

98
00:08:22,180 --> 00:08:25,320
Will lose out in the case of inequality.

99
00:08:26,400 --> 00:08:27,020
Room them.

100
00:08:28,680 --> 00:08:31,140
The jet, which is 30 turps ratio.

101
00:08:32,250 --> 00:08:35,910
Or, Rob, it is the proportion of our population.

102
00:08:37,280 --> 00:08:38,720
And evidence distance.

103
00:08:41,860 --> 00:08:44,950
You can say that these are the most significant variables.

104
00:08:45,160 --> 00:08:45,970
What a what more than.

105
00:08:47,200 --> 00:08:52,090
These are the variables which have the highest impact on the worldwide variable.

106
00:08:53,400 --> 00:08:55,680
So how to interpret this data?

107
00:08:56,790 --> 00:09:00,150
First, you should look at the sign of this patient and.

108
00:09:01,350 --> 00:09:08,370
If the sign is positive, that means if you're increasing your independent variable, your dependent

109
00:09:08,370 --> 00:09:10,350
variable will increase with it.

110
00:09:10,950 --> 00:09:12,000
So, for example.

111
00:09:13,830 --> 00:09:21,120
I'm on the looking at the significant variable, for example, this room, non variable, the coefficient,

112
00:09:21,120 --> 00:09:21,930
dispositive, positive.

113
00:09:23,300 --> 00:09:25,200
And the value of coefficient is for.

114
00:09:26,540 --> 00:09:34,580
So in a way, I'm seeing that if I am increasing are growing numb by one, my price is going to increase

115
00:09:34,580 --> 00:09:35,690
by 40 units.

116
00:09:36,960 --> 00:09:45,480
So in a way, you can say that if all other variables are consent for two houses, if one of the house

117
00:09:45,900 --> 00:09:51,970
has more room numbers than the other house, then the price is going to be 40.

118
00:09:51,990 --> 00:09:53,190
Onex more.

119
00:09:56,040 --> 00:09:58,470
Now, let's look at the air quality here.

120
00:09:58,500 --> 00:09:59,670
My sign is negative.

121
00:10:00,750 --> 00:10:06,840
So in a way, I'm saying that if we are increasing the value of air quality, the price is going to

122
00:10:06,840 --> 00:10:07,500
decrease.

123
00:10:10,810 --> 00:10:15,280
Now, the p value for airport variable is also very less.

124
00:10:15,400 --> 00:10:17,530
That is zero point zero one three.

125
00:10:19,970 --> 00:10:22,400
And here the coefficient is one point one.

126
00:10:23,600 --> 00:10:30,740
So in a way, we're seeing that if the airport is present in any city, the prize is going to be increased

127
00:10:30,740 --> 00:10:31,670
by one unit.

128
00:10:32,870 --> 00:10:39,710
We are also somewhat confident that this variable have a significant effect on our dependent variable.

129
00:10:41,870 --> 00:10:50,570
So important takeaways from this research is identify the variables which have P value less than zero

130
00:10:50,570 --> 00:10:51,560
point zero five.

131
00:10:52,280 --> 00:11:00,230
And try to observe their coefficient and their signs to make business sense out of your data.

132
00:11:02,730 --> 00:11:03,630
This is the desert.

133
00:11:03,720 --> 00:11:08,310
From where you will Assad be driving your business intelligence?

134
00:11:09,400 --> 00:11:15,310
So in this case, you can say that you had house price is dependent on this five or six variables,

135
00:11:15,430 --> 00:11:24,280
which is room air quality and so on, and you can also tell that all this variables are impacting your

136
00:11:24,280 --> 00:11:24,880
price.

137
00:11:27,090 --> 00:11:32,040
Now, we will also create this similar model using a Escalon library.

138
00:11:34,610 --> 00:11:37,820
Our X and Y variables are ready with us.

139
00:11:38,030 --> 00:11:40,190
So it will just create another object.

140
00:11:41,400 --> 00:11:44,240
L m three l m three.

141
00:11:44,300 --> 00:11:50,480
Is the word object equal to linear regression?

142
00:11:58,860 --> 00:12:03,760
Alem three is our variable and linear regression is a function of Escalon Library.

143
00:12:03,940 --> 00:12:05,740
We have already imported Escalon.

144
00:12:07,330 --> 00:12:10,200
Now, to fit this model we let them through don't fit.

145
00:12:11,770 --> 00:12:14,880
Now our X was X underscore my D..

146
00:12:15,640 --> 00:12:21,820
We don't need another variable with the consent sign in case of Escalon.

147
00:12:21,910 --> 00:12:23,830
So we'll just write a Eskander score.

148
00:12:25,430 --> 00:12:29,730
Might be another way variable this way, underscore maybe.

149
00:12:40,560 --> 00:12:44,050
We have for today, at least there, more done with a.

150
00:12:45,340 --> 00:12:52,170
Now, to get the coefficient, we will just, I think, Alem three.

151
00:12:54,750 --> 00:12:57,750
Don't intercept the school.

152
00:13:04,600 --> 00:13:06,850
Coma three.

153
00:13:08,130 --> 00:13:17,840
Well, if endoscope intercept underscored attributes represent we dunnart and quiff underscore a tribute

154
00:13:17,840 --> 00:13:21,390
group resing we dulan with our group with the three other.

155
00:13:23,480 --> 00:13:26,880
You can see first we have the value of intercept.

156
00:13:27,350 --> 00:13:30,270
And then we have the value of be done with that, too.

157
00:13:31,040 --> 00:13:33,740
You can also compare this result with the.

158
00:13:34,860 --> 00:13:37,980
Result we got from Start Smardon.

159
00:13:42,980 --> 00:13:43,760
You can see.

160
00:13:44,810 --> 00:13:48,680
It's not possible to get at somebody of a Linnean more than in a Skillern.

161
00:13:49,700 --> 00:13:57,020
But Escalon is much more useful and on other aspects by fitting the model and while predicting the values

162
00:13:57,170 --> 00:14:00,380
and running other advanced linear regression models.

163
00:14:02,050 --> 00:14:06,230
That's why we are also teaching you how to run this Eskinder.

164
00:14:09,020 --> 00:14:12,860
That is how we run a building integration, Martin and Biton.
