Additional file 4.Goodness of fit for multiple linear regression. Estimates of the square root of residual variance, σ, are reported for each time point and were calulated by the MATLAB function robustfit in order to aggregate the residuals into a single measure of predictive power. First, a σ estimate (root-mean-square-error) is calculated from ordinary least squares (σ_{OLS}), and a robust estimate of sigma (σ_{robust}) is also calculated. The final estimate of σ is the larger of σ_{robust }and a weighted average of σ_{OLS }and σ_{robust}. Note that σ is equal to median absolute deviation (MAD) of the residuals from their median, scaled to make the estimate unbiased for the normal distribution: σ = MAD/0.6745. Also shown are the mean of the residuals at each time point. To put residuals on a comparable scale, they are "studentized," that is, they are divided by an estimate of their standard deviation that is independent of their value. Format: PDF Size: 269KB Download file This file can be viewed with: Adobe Acrobat Reader Rao and Pellegrini BMC Systems Biology 2011 5:160 doi:10.1186/1752-0509-5-160 |