-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathen_Evaluation Metrics in Regression Models.srt
160 lines (120 loc) · 3.83 KB
/
en_Evaluation Metrics in Regression Models.srt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
0
00:00:00,680 --> 00:00:02,730
Hello, and welcome!
1
00:00:02,730 --> 00:00:07,299
In this video, we’ll be covering accuracy metrics for model evaluation.
2
00:00:07,299 --> 00:00:10,810
So, let’s get started.
3
00:00:10,810 --> 00:00:14,620
Evaluation metrics are used to explain the performance of a model.
4
00:00:14,620 --> 00:00:19,860
Let’s talk more about the model evaluation metrics that are used for regression.
5
00:00:19,860 --> 00:00:26,320
As mentioned, basically, we can compare the actual values and predicted values to calculate
6
00:00:26,320 --> 00:00:29,970
the accuracy of a regression model.
7
00:00:29,970 --> 00:00:35,230
Evaluation metrics provide a key role in the development of a model, as it provides insight
8
00:00:35,230 --> 00:00:37,820
to areas that require improvement.
9
00:00:37,820 --> 00:00:42,579
We’ll be reviewing a number of model evaluation metrics, including:
10
00:00:42,579 --> 00:00:48,469
Mean Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE).
11
00:00:48,469 --> 00:00:55,129
But, before we get into defining these, we need to define what an error actually is.
12
00:00:55,129 --> 00:01:00,170
In the context of regression, the error of the model is the difference between the data
13
00:01:00,170 --> 00:01:04,449
points and the trend line generated by the algorithm.
14
00:01:04,449 --> 00:01:10,470
Since there are multiple data points, an error can be determined in multiple ways.
15
00:01:10,470 --> 00:01:15,770
Mean absolute error is the mean of the absolute value of the errors.
16
00:01:15,770 --> 00:01:21,600
This is the easiest of the metrics to understand, since it’s just the average error.
17
00:01:21,600 --> 00:01:25,250
Mean Squared Error (MSE) is the mean of the squared error.
18
00:01:25,250 --> 00:01:31,979
It’s more popular than Mean absolute error because the focus is geared more towards large
19
00:01:31,979 --> 00:01:33,200
errors.
20
00:01:33,200 --> 00:01:38,619
This is due to the squared term exponentially increasing larger errors in comparison to
21
00:01:38,619 --> 00:01:40,539
smaller ones.
22
00:01:40,539 --> 00:01:43,050
Root Mean Squared Error (RMSE) is the square
23
00:01:43,050 --> 00:01:46,270
root of the mean squared error.
24
00:01:46,270 --> 00:01:52,159
This is one of the most popular of the evaluation metrics because Root Mean Squared Error is
25
00:01:52,159 --> 00:01:59,310
interpretable in the same units as the response vector (or ‘y’ units) making it easy to relate
26
00:01:59,310 --> 00:02:01,409
its information.
27
00:02:01,409 --> 00:02:08,560
Relative Absolute Error (RAE), also known as Residual sum of square, where y-bar is a mean value
28
00:02:08,560 --> 00:02:16,409
of y, takes the total absolute error and normalizes it by dividing by the total absolute error
29
00:02:16,409 --> 00:02:18,920
of the simple predictor.
30
00:02:18,920 --> 00:02:24,310
Relative Squared Error (RSE) is very similar to “Relative absolute error “, but is widely
31
00:02:24,310 --> 00:02:30,489
adopted by the data science community, as it is used for calculating R-squared.
32
00:02:30,489 --> 00:02:37,219
R-squared is not error, per se, but is a popular metric for the accuracy of your model.
33
00:02:37,219 --> 00:02:42,909
It represents how close the data values are to the fitted regression line.
34
00:02:42,909 --> 00:02:47,900
The higher the R-squared, the better the model fits your data.
35
00:02:47,900 --> 00:02:52,250
Each of these metrics can be used for quantifying of your prediction.
36
00:02:52,250 --> 00:02:58,359
The choice of metric completely depends on the type of model, your data type, and domain
37
00:02:58,359 --> 00:02:59,879
of knowledge.
38
00:02:59,879 --> 00:03:04,499
Unfortunately, further review is out of scope of this course.
39
00:03:04,499 --> 00:03:05,589
Thanks for watching!