Fig. 1

Conceptual Figure of the Proposed Algorithm. A test case of the algorithm using a simulated movement trajectory of 1000 relocation points (a). Three of the subsets of those points, with red points indicating those locations that remain in the training sets and blue points representing the test points for the later probability calculation (b,f,j). For each subset of points, a hullset is created using T-LoCoH, with an arbitrarily chosen s value of 0.5 and k values of 5 (c,g,k), 15 (d,h,l), or 25 (e,i,m). These three subsets serve to illustrate three possible scenarios as the k values increases: either test points that are not covered by the hull set at low k values continue to be uncovered with high k values (left-hand column), test points that were not originally covered by the hull set at smaller k values becomes covered (center column), or test points are covered at low k values and continue being covered at higher k values (right-hand column)