Package ai.kognition.pilecv4j.nr

Class SimpleLinearRegression

java.lang.Object
ai.kognition.pilecv4j.nr.SimpleLinearRegression
All Implemented Interfaces:
Minimizer.Func
public class SimpleLinearRegression extends Object implements Minimizer.Func

This class will do a linear regression by minimizing the squared error between the points provided to the constructor and the line specified by y = m[0]x + m[1]

NOTE: The error term is calculated using the vertical distance. That is it minimizes the error in the 'y' term only. This is more robust for determining the slope/intercept form of a line because the starting point cannot be rotated through the Y axis. In other words, if the slope of the points is approximately -3.0 and the slope of the starting iteration is 3.0, then RAISING the slope lowers the perpendicular distance. This moves the iterations in the wrong direction.

  • Field Details

    • y

      public final double[] y

    • x

      public final double[] x

  • Constructor Details

    • SimpleLinearRegression

      public SimpleLinearRegression(double[] x, double[] y)

  • Method Details

    • func

      public double func(double[] lineDefMb)
      Specified by:
      func in interface Minimizer.Func