Figure 4.

Example of a simulation in the irregular case, and result of the dynamic programming algorithm for the estimation of the break-point coordinates. Top: Example of simulation for 100 data points and 5 segments in the irregular case. The true break-points are designated by vertical lines, and the red lines correspond to the mean of each segment. The difference of means varies between d = 2 to d = 0.5. Bottom: Estimated probability for a break-point to be located at coordinate t for t = 1 to 100. Different levels of noise are considered with σ = 0.1, σ = 0.5, σ = 1.

Picard et al. BMC Bioinformatics 2005 6:27   doi:10.1186/1471-2105-6-27
Download authors' original image