Summary
Although the main purpose of iSAX is to help us search for subsequences by indexing them, there are other ways to use an iSAX index.
In this chapter, we presented a way to approximately compute the Matrix Profile vectors and two ways to approximately compute the MPdist distance between two time series. All these techniques use iSAX indexes.
We presented two ways to approximately compute MPdist. Out of the two methods, the one that joins two iSAX indexes is much more efficient than the other – so the use of an iSAX index by itself does not guarantee efficiency; we have to use an iSAX index the right way to get better results.
There is a small chapter left to finish this book, which is about the next steps you can follow if you are really into time series and databases.