Data Science Capstone Final Project

• Course: Data Science Specialization Track-Data Science Capstone Project
• Provider: Coursera @ Johns Hopkins University
  
• Goal: Build an interactive web app for next-word prediction
• Technology: Shiny (R) + Distributed N-gram Language Model
• Purpose: Predict the next word in a user-typed phrase

Distributed N-gram Language Model with Back-off Logic

Training Data: Twitter, Blogs, and News corpora

How it works: If no quadgram match exists, the model recursively searches trigrams, then bigrams (back-off logic)

What the App Does

• Type a phrase → N-gram model suggests what comes next
• Displays user input with live analysis
• Shows 1-3 recommended word suggestions (configurable)
• Uses quadgram back-off for higher accuracy

How It Functions

1. User enters text in the Text Input area
2. Selects Suggestion Count (1-3 words) using slider
3. Clicks Predict Next Word button
4. App processes input through N-gram model
5. Top predictions displayed as blue “chip” buttons with loading spinner

How to Use the App

1. Navigate to the Home tab
2. Type a phrase (e.g., “How are you”)
3. Adjust suggestion count using slider (1-3)
4. Click “Predict Next Word” button
5. View predicted words in blue chips

• Shiny App link : https://swarup-jhu.shinyapps.io/Data-Science-Capstone-Final-Project/
• Github Link : https://github.com/swarup-jhu/Data-Science-Capstone-Final-Project