Reminders: for phase 3 please include
1) problem definition
2) list of your innovations
Here are some examples
1) Precise PROBLEM DEFINITION - like, eg.,
Given:
* N time sequences, of equal duration $T$
* a query sequence $s$ of same duration
* a user-defined distance function D(s1,s2) between two sequences s1,s2
Find
* the k-nearest-neighbors to query sequence $s$
* in sub-linear time (< O(N) )
2) List of your INNOVATIONS, like, e.g.,
* Our method uses ,
thus being faster than all the competitors we surveyed.
* Scalability: our method scales sub-linearly with $N$, the number
of sequences in the database
* Correctness: we proved that our method has no false dismissals
and thus it returns exactly the same results as sequential scan
christos