Successfully automated curve fitting is greatly challenged when applied to large data set. In this paper, we described a algorithm for fitting dose response curves, by estimating four parameters (floor, window, shift, and slope), together with the det...
Successfully automated curve fitting is greatly challenged when applied to large data set. In this paper, we described a algorithm for fitting dose response curves, by estimating four parameters (floor, window, shift, and slope), together with the detection of outliers. Especially, We are proposing an improvement for curve fitting over current methods. That is the detectionof outliers which is performed at the initialization step with correspondentadjustments of the derivative and error estimation functions. Automatic curve fitting of 19,236 experimental dose response experiments shows that our approach outperformed the current fitting methods provided by Matlab nlinfit function.