Can anyone point me to a simple linear regression algorithm. Nothing fancy, just best fit linear regression line for a set of X/Y coordinates using least squares. C# preferred, but I'll take what I can get.
Thanks in advance.
Sean
Printable View
Can anyone point me to a simple linear regression algorithm. Nothing fancy, just best fit linear regression line for a set of X/Y coordinates using least squares. C# preferred, but I'll take what I can get.
Thanks in advance.
Sean
If you are simply looking for the regression equation, you can use the following formulae:Quote:
Originally Posted by sean_colonello
y = b1*x + b0
b1 = sum((x_i-x_avg)*(y_i-y_avg)) / sum((x_i-x_avg)^2)
b0 = y_avg - b1*x_avg
If you also want other statistical data (such as standard deviation, correlation, etc.) there are other, more efficient ways to acquire the data since they are all related. But you don't need an algorithm, just an equation.