Problem: Users often need help predicting the next word while typing.
Goal: Build a Shiny app that predicts the next word in a phrase using n-gram modeling.
Audience: Anyone who wants to speed up typing or autocomplete phrases.

• Approach: N-gram backoff model
  • Trigrams: Predict next word using last two words
    
  • Bigrams: If trigram not found, use last word
    
  • Unigrams: If nothing found, use most common word
    
• Text Cleaning: Convert to lowercase, remove punctuation, remove extra spaces
  
• Data: Sample training data of English phrases

• Input: User types a phrase in a text box
  
• Action: Click “Predict Next Word”
  
• Output: Predicted next word is displayed
  
• Implementation: Shiny app using R, prediction.R logic, ngrams.RData data

• Simple and clean interface
  
• Works for multiple example phrases:
  • data science is → powerful
    
  • the stock market → is
    
  • one of the most → important
    
• Provides quick, accurate prediction

• Novelty: Small, fast n-gram backoff model for real-time predictions
  
• Scalability: Can extend to larger corpora for better accuracy
  
• Conclusion: User-friendly, functional app ready to demonstrate predictive text capabilities