Typing on mobile devices can be difficult and slow.

• Slow Speed: Typing long sentences takes time.
• Typos: Small screens lead to frequent spelling errors.
• The Need: We need a smart algorithm that can predict the next word to speed up typing.

I have developed a Shiny App that predicts the next word based on user input.

• User Input: The user types a phrase (e.g., “Good”).
• Processing: The app cleans the text and checks the database.
• Output: It instantly suggests the most likely next word (e.g., “Morning”).

The app uses a Backoff N-Gram Model built on a large dataset of Twitter, News, and Blogs.

1. Trigram Search: First, it looks for a match in 3-word combinations (High Accuracy).
2. Bigram Search: If no match is found, it backs off to 2-word combinations.
3. Default: If nothing is found, it predicts the most common word (“the”).

The application is designed for speed and simplicity.

• Clean Interface: Just a simple text box and a result display.
• Fast Performance: Uses data.table for sub-second predictions.
• Mobile Ready: Lightweight and works on any device.

This Data Science Capstone project demonstrates the power of Natural Language Processing (NLP).

• Accurate: Uses real-world data patterns.
• Efficient: Optimized for low memory usage.
• Scalable: Can be improved with more data.

Thank You!