Demo of Simple Linear Least Squares Regression and 3 Other Fitting Methods скачать в хорошем качестве

Demo of Simple Linear Least Squares Regression and 3 Other Fitting Methods

This is a demonstration using least squares, least sum of absolute residuals regresion, and least median squared residual regression. The latter two are more robust to outliers. All three approaches were forced to fit straight lines here. A fourth method that does not assume linearity, the loess nonparametric smooter, is also demonstrated. The R code for the demo may be found at https://github.com/harrelfe/rscripts/.... It can be run under RStudio (with clicked data points only showing for an instant until data entry is complete) using the R commands require(Hmisc); getRs('demoLeastSquares.r') or under plain R using the commands require(Hmisc); getRs('demoLeastSquares.r', put='source').