PRESEE: An MDL/MML Algorithm to Time-Series Stream Segmenting

Algorithm 2

PRESEE.

Input: windowSize, S = ⟨𝑠₁,𝑠₂⋯〉, batchSize

Output:  C = ⟨𝑐₁,𝑐₂⋯⟩

(1)  slideWindow = {}, apprSet = {}

(2)  Read data into slideWindow

(3)  While  ReadIn() do

(4)   tmpSet = MDLSlideWindow(slideWindow,

    batchSize);

(5)   If  tmpSet.size() < 2 do

    // no proper hypothesis is found,

      data deemed as random noise

(6)    Add all data in slideWindow into apprSet

(7)    Empty slideWindow

(8)    Else

(9)    Add at most first batchSize + 1 points in

     tmpSet into apprSet

(10)       Take out used points from slideWindow

(11)       Output apprSet