regression is fitting line onto data points
Table of Contents
The model is known in regression analysis (e.g. y = kx + b), and with regression algorithms, parameters (e.g. k and b) would be approximated. As result, a real line, instead of a parameterized (e.g. with k = 3 and b = -1) can be drawn, fitting the data point, providing clear correlation insights
Backlinks
regression
Knowing:
- data points \(\vec x = (x_1,x_2,...,x_n) \in X\)
- \(\vec x = (x_1...x_n)\) satisfy \(f(\vec x,P) = f(x_1,x_2,...,x_n,P) = 0\) with a set of parameters \(P\)
Find:
- \(P\) (which due to noise in \(X\) is unlikely) or
- \(P'\) that optimize/minimize some function \(f'(\vec x,P')\) (usually error of \(f(\vec x,P')\) over \(X\))
- regression is fitting line onto data points
- linear regression
- polynomial regression
- gnuplot line fitting