Overview
This project builds a next-word prediction model using a bigram
language model.
The application is deployed using Shiny and provides real-time word
prediction.
Algorithm
- Text preprocessing and tokenization
- Bigram frequency model
- Frequency-based next word selection
- Default fallback word if no match found
Application
- User enters a phrase
- App extracts last word
- Searches bigram model
- Returns predicted next word
User Experience
- Simple and fast interface
- Instant word prediction
- Interactive web deployment
Conclusion
- Lightweight predictive model
- Expandable to higher n-grams
- Deployable as scalable web application