next up previous contents
Next: Representing time explicitly Up: New heuristics Previous: New heuristics

Temporal Concurrency and Sequences

 

If events occur concurrently across different channels, and this occurs consistently in our training data, then there is a high probability that this is not a coincidence and can be considered as a defining characteristic of the class. Similarly, if a particular sequence of events occurs consistently for a given class, then our learning algorithm can take advantage of this.



Mohammed Waleed Kadous
Tue Oct 6 13:04:40 EST 1998