Question 1190381: A data series is given for a periodic process:
0 1 2 3 4 5 6 7 8 9 10
1, 7 1, 7 2, 3 2, 9 3, 0 3, 5 4, 0 11, 0 9, 5 9, 7 9, 9
Please find a linear regression.
Is the use of a linear regression okay
Answer by math_tutor2020(3816)   (Show Source):
You can put this solution on YOUR website!
The data table seems a bit garbled.
This is what I'm assuming it looks like
| 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | | (1,7) | (1,7) | (2,3) | (2,9) | (3,0) | (3,5) | (4,0) | (11,0) | (9,5) | (9,7) | (9,9) |
You could find the linear regression equation by hand, but it's a tedious process that may take a while.
I strongly recommend using a graphing calculator or computer software. There are many free options online if you don't have a graphing calculator.
Whichever method you use, the linear regression equation is approximately this:
y = -0.0929x + 5.1832
The correlation coefficient (r value) is a similar story to finding the regression equation. It's better to use a calculator.
You should find that r = -0.1004 approximately
Since this is much closer to r = 0 than it is to r = 1, this makes the negative linear correlation fairly weak.
It's not a good idea to use a linear model here.
Because you mentioned the data is periodic, a sine or cosine model would likely provide a better fit.
|
|
|
