Cluster an array of patterns with an overlap delta This is with a streaming window > 1 A pattern would then comprise of rows of data concatenated together where the number of rows concatenated is equal to the step count This functions as a shifting window where a step count of 3 would take rows 1 through 3 for the first pattern, 2 through 4 for the second pattern, 3 through 5 for the thrid pattern, etc
- Parameters
-
| PatternBuffer | Buffer where 2 or more patterns are stored contiguously |
| NumPatterns | The number of patterns found in PatternBuffer |
| ResultBuffer | Buffer of length NumPatterns where array element ResultBuffer[i] will get the cluster ID assigned to pattern i in PatternBuffer |
| StepCount | Number of consecutive rows of data that will make up one pattern |
| AdvancedAnalyticsBuffer | If non null, will be populated by the cluster statistics analytics for each cluster ID returned in the ResultBuffer |
- Returns
- The number of patterns that were generated from the data given the step count