WordSmart is a web based app created to predict the next word of a given phrase or string of text.

WordSmart helps the user to type in a fun and efficient way by using a machine learning algorithm previously trained.

The algorithm to predict the next word has been optimized to be fast and reliable. Some characteristics of the WordSmart prediction model:

• The model relies on N-Gram modeling used frequently by NLP applications.

• The model exploits several features provided by the Quanteda library.

• The model uses Markov chains, which state that the probability of the next possible word depends on the last word provided.

• The model returns unigrams with the highest probability of occurring next.

1. Open the WordSmart app

2. Type a phrase in the text input

3. Click the Predict button

• Wordsmart is fast and reliable
• Easy to use
• Provides several suggestions
• Trained on a large corpus of text data provided by Swiftkey
• Useful for many applications in which word prediction is useful.