Introduction

In many real-world applications, users need text prediction for fast communication or assistance. The solution is to have a Shiny app powered by an n-gram model (unigram, bigram, trigram) to predict the next word based on user input. So the objective of this app is to present an interactive predictive text model that enhances user experience in writing and typing applications. The app can be found here : https://acecube.shinyapps.io/Text_Generator/

How the Algorithm Works

• Input: The user provides a piece of text.
• Algorithm:

1. Trigram Prediction: First, check if the last two words can predict the next word using a trigram model.
2. Bigram Prediction: If no trigram match is found, fall back to the bigram model using the last word.
3. Unigram Prediction: If no bigram match, use the unigram model for a fallback prediction.

• Output: The app predicts the next word based on the input and model matches.

Key Features

• User Input: Users type a sentence or phrase in the text box.
• Instant Prediction: The app instantly predicts the next word when the button is clicked.
• Model Flexibility: The app uses different levels of n-grams for increasingly accurate predictions.
• Technology: Built using R, Shiny, and the dplyr package for data manipulation.

User interface

• App Interface: The user interface consists of:

1. A text input box for typing.
2. A button that triggers the word prediction.
3. An area to display the predicted next word.

• Interactive Design: Users can interact with the app in real-time to test predictions for any given input.

Potential Applications & Benefits

• Text Assistance: Perfect for typing applications, chatbots, or email assistants.
• Increased Efficiency: Saves time in writing and enhances productivity.
• Customizable: Easily adaptable for other languages or specialized domains (e.g., medical or legal).
• Market Potential: Can be scaled for commercial use in various software and services.