Problem

• Typing on mobile devices is slow and error-prone.
• Predictive typing helps users type faster and more accurately.
• Goal: Build a Shiny app to predict the next word based on typed input.

The Algorithm

• Based on N-gram models (Bigrams & Trigrams).
• Trained on SwiftKey dataset (Twitter, Blogs, News).
• Steps:
  • Clean input
  • Extract last 1–2 words
  • Search N-gram table
  • Return most frequent match

The Shiny App

• Built with R Shiny
• Simple UI:
  • User types a phrase
  • App predicts the next word
• Deployed here:
  https://yourusername.shinyapps.io/textpredictor/
  (Replace with your actual link)

Value Proposition

• Increases typing speed by 30–50%
• Can be adapted for:
  • Mobile keyboards
  • Chatbots
  • Accessibility tools
• Lightweight, responsive, and easy to scale

Future Plans

• Add personalization based on user history
• Explore transformer-based models
• Enable multilingual prediction
• Commercial licensing to messaging & keyboard apps