Checking the search speed of iSAX indexes

This section presents a utility that takes two time series, named TS1 and TS2, which ideally have similar lengths, creates two iSAX indexes, named D1 and D2, and performs the following searches:

• Searches D1 for all the subsequences of TS2. In this case, we are not sure whether a subsequence from TS2 is in D1 or not. In most cases, we are not going to be able to find the subsequences of TS2 in TS1. This is the main reason that a join based on the SAX representations of the iSAX nodes might be more appropriate when looking for similarities between subsequences.
• Searches D2 for all the subsequences of TS1. In this case, we are not sure whether a subsequence from TS1 is in D2 or not. As before, in most cases, we are not going to be able to find the subsequences of TS1 in TS2 and therefore, in the iSAX index created from TS2 (D2).
• Searches D1 for all the subsequences of TS1, which means that all subsequences of TS1 are in D1. With...