Research Article
Prediction after a Horizon of Predictability: Nonpredictable Points and Partial Multistep Prediction for Chaotic Time Series
Algorithm 1
The Wishart clustering algorithm.
| | determine = distance to the sample’s -nearest neighbor; | | | sort in ascending order; | | | q = 1; | | | for each subgraph : | | | while : | | | = newly added vertex of the subgraph; | | | if is not connected to any clusters: | | | start new cluster; | | | else: | | | if connected to the vertices of clusters : | | | if all clusters are completed: | | | = 0; | | | else: | | | = number of significant clusters; | | | if or = 0: | | | = 0; | | | label significant clusters as completed; | | | delete labels of insignificant clusters; | | | else: | | | merge clusters into ; | | | = ; | | | set for samples in ; | | | q = q + 1; |
|