1. Create corpus and clean data with tm
2. Create bi-, tri- and tetragrams with RWeka
3. In server.R (shiny)
   • Strip user input to last three (two, one) words
   • Search first three words of tetragrams
   • If no matches, search first two words of trigrams
   • If no matches, search first word of bigrams
   • Calculate percentages for matches

Performance

Less than five seconds
response time

The first search
might take a bit longer
when the app is awakening

Visit the