2010-10-20, 10:33 PM
I have the following data:
x l y
0.5 33.3
1.0 58.3
1.5 81.8
2.0 96.7
2.5 100.0
I am told to either use the square root of x or log(x) instead of x. I use the square root of x as my transformation, replace it with x, and find the equation of the original equation using a linear reg on the transformed data. Here's the problem: When I plug 1.75 into my calculated linreg equation, I get ~23.3, which is wayyy offf of the actual value (I put the equation in the form of y=a+b(rootx). Did I do something wrong, or did I just use the wrong transformation?
x l y
0.5 33.3
1.0 58.3
1.5 81.8
2.0 96.7
2.5 100.0
I am told to either use the square root of x or log(x) instead of x. I use the square root of x as my transformation, replace it with x, and find the equation of the original equation using a linear reg on the transformed data. Here's the problem: When I plug 1.75 into my calculated linreg equation, I get ~23.3, which is wayyy offf of the actual value (I put the equation in the form of y=a+b(rootx). Did I do something wrong, or did I just use the wrong transformation?