A Shiny App powered by Bigram Prediction Algorithm

• Built in R & Shiny
  
• Predicts the next word based on a phrase
  
• Deployable on shinyapps.io

• Typing is slow and repetitive in many applications.
  
• Predictive text can help users save time by suggesting the next word.
  
• Our goal:
  • Build a lightweight Shiny app to predict the next word in real-time.
    
  • Create a prototype that can be scaled to real-world applications. ## Prediction Algorithm

Bigram Model (Statistical Language Model)

1. Training:
   • Break sentences into word pairs (bigrams).
     
   • Count how often each pair occurs.
2. Prediction:
   • Take the last word from the user’s phrase.
     
   • Suggest the most frequent word that follows in the training data.

Example:

Input: “I am” → Bigram Table: (“am”, “learning” → 5 times, “am”, “working” → 3 times) Predicted Word: “learning”

• How it Works:
  1. Enter a phrase (e.g., “I am”).
     
  2. Click “Predict Next Word”.
     
  3. The app shows the predicted next word.
• User Interface:
  • Simple input box + one-click prediction.
    
  • Fast and lightweight.
• Experience:
  • Easy to use even for non-technical users.
    
  • Real-time response, feels like using mobile keyboard prediction.

• Applications:
  • Smart keyboards, chatbots, auto-complete in emails.
    
  • Writing assistants & customer support bots.
• Future Enhancements:
  • Use Trigram / Deep Learning (LSTM, Transformers) for better accuracy.
    
  • Train on real-world large datasets.
• Why It’s Novel?
  • Lightweight, deployable in a browser.
    
  • Can be expanded to a full predictive typing system.

Would you invest? → Yes, because predictive text is a billion-dollar market!

Thank You