Wavelet Trees - an Introduction

13 years ago (alexbowe.com)

Sounds interesting, but theoretical. Any practical applications? I'm guessing it has some purpose on dealing with long string sequences, like genome sequencing?

  • From what I understand, it is useful for storing compressed data and querying it without decompression.

    Range queries allow for full-text search, for example.

  • Something like that + data compression. If you have some data type mapped to a sequence of symbols, WT can speed up search in it. Please also note it is not a complete full text search index (like ones used in genome sequencing). It is a rank/select dictionary.

    • I believe he's doing full text search on this by running Burrows-Wheeler on the data first. See here for more details:

         http://alexbowe.com/fm-index/
      

      I started reimplementing all of this in Go about a year ago, but moved onto other stuff before I finished.

      1 reply →

I would love to see more algorithms from the signals intelligence field. Any recommendations?