Text prediction shiny application created for Coursera-Johns Hopkins University

Alfredo Aranda
August 08th, 2019

Predictions

The aim of this application is to meet the need of a text predictive Shiny application which can predict the users next word with speed and accuracy with the emphasise on the latter.

“It will not matter how fast the app is, if it is inaccurate.”

The algorithm is based on a N-Gram model that was built from a large corpora supplied by SwiftKey. The material was sourced from Twitter, News and Blogs.

Corpora Stats

Total Size Total Words
28MB 4,500,000

Shiny App Interface

Woracle uses a clean minimal user interface (UI) which access the data compression and predict algorithems working in the background.

The user enter text into the input box and waits for the next word to be predicted.

Algorithm behind the app

Fast and simple, the algorithm is build for accurancy first, then speed. The average typist types 36WPM, the algorithm will work to meet that critiria.

Trigam model was utilized.

  • accurancy: 60%

  • speed: 1 to 2 secs

  • memory used: 4MB

  • hard disk used: 50MB

Try Woracle Today

Enter a sentence or a phrase in the box and you can get a prediction of the next word!

Thanks!