RANSAC polyfit. Fit polynomials with RANSAC in Python

ransac_polyfit.py

def ransac_polyfit(x, y, order=3, n=20, k=100, t=0.1, d=100, f=0.8):
  # Thanks https://en.wikipedia.org/wiki/Random_sample_consensus
  
  # n – minimum number of data points required to fit the model
  # k – maximum number of iterations allowed in the algorithm
  # t – threshold value to determine when a data point fits a model
  # d – number of close data points required to assert that a model fits well to data
  # f – fraction of close data points required
  
  besterr = np.inf
  bestfit = None
  for kk in xrange(k):
    maybeinliers = np.random.randint(len(x), size=n)
    maybemodel = np.polyfit(x[maybeinliers], y[maybeinliers], order)
    alsoinliers = np.abs(np.polyval(maybemodel, x)-y) < t
    if sum(alsoinliers) > d and sum(alsoinliers) > len(x)*f:
      bettermodel = np.polyfit(x[alsoinliers], y[alsoinliers], order)
      thiserr = np.sum(np.abs(np.polyval(bettermodel, x[alsoinliers])-y[alsoinliers]))
      if thiserr < besterr:
        bestfit = bettermodel
        besterr = thiserr
  return bestfit