Comparisons of Split-linear Fitting of Wind Curves

Abstract

The detection of slope change points in wind curves is associated with linear curve tting. Classic algorithms based on smoothing and comparison of slopes are adapted and compared to a Bayesian method of curve tting. After prior spline smoothing of the data, the algorithms are tested and the errors between the split-linear rtted wind and the real wind are estimated. In our case, adaptations of an edge-preserving smoothing algorithm allow the same good performance as automatic Bayesian curve etting based on a Monte Carlo Markov chain algorithm.