Archived; click post to view.
Excerpt: Over one year ago I was starting to think about indexing scored stings for auto completion queries. I stumbled upon this problem after seeing the strength of the predictions of the typology approach for next word prediction on smartphones. The typology approach had one major drawback: Though its suggestions had a high precision the speed with 50 milliseconds per suggestion was rather slow especially when working on a server side application. On August 16 th 2012 I found a first solution building on Nicolai Diethelms Suggest Tree. Though the speedup was great the suggest tree at this time had several major…
Related posts:
- Building an Autocomplete Service in GWT screencast Part 3: Getting the Server code to send a basic response In this screencast of my series on building an autocomplete...
- Create a Screencast in Ubuntu with recordmydesktop and do Soundengineering and post production I promised to create some screen casts for the autocomplete...
