1
00:00:00,390 --> 00:00:08,430
In this video, we will assess the accuracy of the model that we have created, that we have established

2
00:00:08,430 --> 00:00:17,040
the relationship between X and Y, but we want to know how well does the predicted values with the actual

3
00:00:17,040 --> 00:00:17,720
Y values.

4
00:00:19,380 --> 00:00:21,270
So assess the quality of it.

5
00:00:21,780 --> 00:00:23,940
We will look at two related qualities.

6
00:00:25,020 --> 00:00:30,000
One is the digital standard edit and the other is called R Square.

7
00:00:32,700 --> 00:00:35,070
Let us first look at the standard error.

8
00:00:38,050 --> 00:00:45,790
We saw earlier that the visitors center that it is on the route are assessed by and minus two, which

9
00:00:45,790 --> 00:00:53,470
can also be written like this, a summation of squared of difference between actual value and predictive

10
00:00:53,470 --> 00:00:53,830
value.

11
00:00:56,130 --> 00:01:02,770
Roughly speaking, Oddisee is the average amount that the response will deviate from the true regression

12
00:01:02,770 --> 00:01:03,030
line.

13
00:01:05,940 --> 00:01:13,270
As you can see in the results shown below, this is the result we got when we ran the model and software,

14
00:01:14,020 --> 00:01:19,170
it is straightaway giving us the standard error, which is six point five minutes of for this model

15
00:01:19,800 --> 00:01:21,700
on 504 degrees of freedom.

16
00:01:22,260 --> 00:01:28,610
This 504 we are getting from minus two and this or to minus two is 504.

17
00:01:29,070 --> 00:01:30,990
And this is called the degrees of Freedom.

18
00:01:32,580 --> 00:01:37,610
So for these many degrees of freedom, we are getting a standard error of six point nine seven.

19
00:01:38,940 --> 00:01:47,580
And in other words, even if the model was correct and true values of zero and we were never known exactly

20
00:01:48,420 --> 00:01:52,410
the predicted value of house price from this model.

21
00:01:53,320 --> 00:01:59,710
They'll still be off by six point five nine seven units on an average.

22
00:02:01,730 --> 00:02:09,650
Therefore, Odyssey can also be considered as a measure of lack of fear of this model to the data.

23
00:02:10,460 --> 00:02:18,470
So this six point five nine seven value is telling you on an average by how many units your predictive

24
00:02:18,470 --> 00:02:21,830
value is missing, the actual value.

25
00:02:25,740 --> 00:02:34,140
Next is the R-squared statistic, The Odyssey provides an absolute measure of lack of it, but since

26
00:02:34,140 --> 00:02:40,650
it is measured in the units of light, it is not always clear what constitutes a good odyssey.

27
00:02:42,870 --> 00:02:45,620
So R-squared provides us with an alternative.

28
00:02:46,590 --> 00:02:48,030
Oscar is a proportion.

29
00:02:49,350 --> 00:02:57,390
The proportion of total variance explained by our model, so it always lies between zero and one, it

30
00:02:57,470 --> 00:03:04,560
is the mathematical formula for Askwith, R-squared, squaddies versus minus Odyssey's upon basis that

31
00:03:04,760 --> 00:03:10,140
assesses total sum of squares and Odyssey's vegetables are Moscowitz.

32
00:03:12,030 --> 00:03:21,360
DSF is measuring the amount of variability inherent in the response that is porras house prices data.

33
00:03:21,720 --> 00:03:25,950
The price of each house itself is varying about the mean house price.

34
00:03:27,110 --> 00:03:34,460
So if you find a difference of actual Osprey's from the man of the house price.

35
00:03:35,520 --> 00:03:44,220
Square these values and add them up, you get total sum of squares, so this total sum of square value

36
00:03:44,400 --> 00:03:48,030
is giving you the total amount of variability in how space.

37
00:03:49,530 --> 00:03:57,450
How much of this is explained by the model that we constructed or that we will use Odyssey's Odyssey's

38
00:03:57,450 --> 00:04:04,290
is measuring the amount of variability that is not explained by our model of the degradation and the

39
00:04:04,350 --> 00:04:05,070
assessed minus.

40
00:04:05,070 --> 00:04:09,120
Odyssey's is giving us the variability of light, which is explained by our model.

41
00:04:10,680 --> 00:04:16,770
Therefore, as good measures, the proportion of explained variance from the total variance.

42
00:04:19,720 --> 00:04:26,200
Oscar winning close to one indicates that a large proportion of the variability in the response variable

43
00:04:26,260 --> 00:04:33,070
has been explained by the regression model, if it is close to zero, it indicates that the regression

44
00:04:33,070 --> 00:04:34,980
did not explain much of the variability.

45
00:04:35,830 --> 00:04:43,510
This can occur either because our linear model is wrong or because linear was not the right choice for

46
00:04:43,510 --> 00:04:50,230
this relationship between X and Y or both of these reasons for our model.

47
00:04:50,800 --> 00:04:56,020
The result given by the software packages that started at about six point forty nine.

48
00:04:57,810 --> 00:05:00,180
R-squared value is zero point forty.

49
00:05:02,470 --> 00:05:09,130
So it is somewhere between zero and one, nearly 50 percent of the variability of the response variable

50
00:05:09,670 --> 00:05:12,250
is handled by the model that reconstructed.

51
00:05:15,490 --> 00:05:21,230
There is another value, we just called it just R-squared, which you can see from the model result,

52
00:05:22,570 --> 00:05:31,270
the difference between this R-squared and this adjusted R-squared is that in adjusted R-squared will

53
00:05:31,270 --> 00:05:36,710
be also taking into account the total number of variables which are actually impacting the model.

54
00:05:38,170 --> 00:05:46,210
The reason behind doing this is if you keep on adding variables to your model, Osgoode value simply

55
00:05:46,210 --> 00:05:47,440
keeps on increasing.

56
00:05:48,890 --> 00:05:56,450
Even if the variable is not significantly related with the response variable, still the Oscar value

57
00:05:56,780 --> 00:05:58,890
will increase by by a small amount.

58
00:06:01,840 --> 00:06:07,750
So the adjusted R-squared is a modified version of R-squared that has been adjusted for the number of

59
00:06:07,750 --> 00:06:15,190
predictors and the model, the adjusted R-squared increases only the new term improved the model more

60
00:06:15,190 --> 00:06:16,210
than would be expected.

61
00:06:16,210 --> 00:06:22,010
By chance, it decreases when the predictor improves the model by less than expected by chance.

62
00:06:23,920 --> 00:06:24,760
So just did.

63
00:06:24,760 --> 00:06:27,370
R-squared is a more preferred term over R-squared.

64
00:06:30,180 --> 00:06:36,560
Now we have a value of R-squared, but what value of R-squared will be considered as a good value of

65
00:06:36,560 --> 00:06:37,020
R-squared?

66
00:06:38,480 --> 00:06:41,920
This will generally depend on the type of application that you have.

67
00:06:42,620 --> 00:06:49,190
If the data is coming from a science experiment and the relationship is supposed to be actually linear

68
00:06:49,790 --> 00:06:52,910
in such a case, Oscar should be very close to one.

69
00:06:54,300 --> 00:07:01,730
But if it is a marketing data and we are missing a lot of unmeasured factors and the linear assumption

70
00:07:01,850 --> 00:07:04,880
is also a rough approximation of the relationship.

71
00:07:06,170 --> 00:07:08,300
These letters are going to be large.

72
00:07:09,730 --> 00:07:17,290
In such a case, even smaller Oscar values can be acceptable generally, Oscar, better than point five

73
00:07:17,440 --> 00:07:19,360
can be considered as a good fit model.