Word Suggestion Solutions

Raymond Bem
January 27, 2021

NextApp Corporation

hyperlink to product…NextApp

Introduction

  • The latest generation of NextApp word suggestion technology has arrived!

  • More Accurate – former reliance on rank-ordered lists results in poor accuracy, maturization of NLP (Natural Language Processing) techniques improve performance considerably

  • Faster – filtering logic means a smaller server footprint as well as faster response time given a language set

Algorithm

  • Highly respected Modified Kneser-Ney Smoothing (MKNS) improves upon traditional back-off search methods

  • Process considers and weights appropriately the word-facings after and before the phrase

  • Lower count phrase options draw insight from the quantity of completion and followup phrases

  • Weigths are calculated based on the frequency of 1 and 2 count phrases, discounting lower level (ngram) search results

  • These discounts reduce errors generally by reducing the effect of highly popular words (e.g., “and”)

App

  • The NextApp is quite simple, the user simply enters a word or phrase and clicks one button

  • Underneath, calibrated tables are built from samples of modern news, blog, and twitter language

  • A final set of weights is applied to the MKNS model probabilities to better surface infrequent phrases

  • The user is returned a list of word selections presented in graphical format – the best choice is clearly called out in the results

  • Distance (y-axis) between the chosen word and other options easily depict strength of selection

Conclusion

  • NextApp Corporation is very excited about this latest work

  • Applications include assisted typing for disabled persons, business inventory consolidation projects, as well as plagiarism detection

  • App can be found at this link…NextApp