December 14, 2014

Featuring Innovative POWN
(Predicting Occurances of Words Numerically) Technology

Proven Design Methodology

  • Markov n-Grams with Backoff Prediction Algorithm
  • Built from 100M Words
  • Real Datasets from the Internet!

Natural Language Generation (NLG)

  • Subset of Natural Language Processing
  • Utilizes Grammars and Statistical Models
  • Extracted from Human Written Texts


now with Context Responsive Usage Design

Our analysis reveals that phrases and word choices are primarily context dependent. Our innovative CRUD system takes context into account when making word choices to reveal subtle changes in various contexts.

Target Your Predictions

  • Prediction Algorithms reorder themselves Dynamically!
  • Choose from any one of News, Blogs and even Twitter Feeds!
  • More Context Feeds can be Added as Needed!

and if that wasn't enough…

Read more of our findings in our detailed report at: []

Behind the Scenes, ICUP is Revealed!

Custom Built, Indexed Compression Utilizing Pointers (ICUP) ensures that your predictions can run on the tiniest of hardware!

Compression Algorithm

  • numerical addresses - virtually elimate space-hogging strings!
  • datasets are loaded individually based on context
  • further reductions in speed and memory requirements are achieved by only requiring one of R's outside libraries. The blazingly fast data.table package
  • we are able to achieve these remarkable results with NO SHINYAPPS SWITCH ENHANCEMENTS

Translation: we are running on the very smallest amount of RAM and Processor speeds that Shinyapps offers!

  1. Choose your Context (News, Blogs or Twitter)
  2. Add your input Phrase - doesn't have to be the whole sentence
  3. Predicted Word shows up on the right
  4. [optional] Click the predicted word to have it added to your phrase and make another prediction
  5. Enjoy! Author: Stephen D. Wells