In this capstone project, I am going to create clone version of the famous SwiftKey app on the phones which will predict the next word given a piece of text.

I used the n-gram method to tackle this problem. By doing so, I will predict the probability of the next words given the sentence input and then return the word that has the most probable within the set of predicted words.

• Like typing on the phone, while you are insert a piece of text on the left-hand side, the app will automatically predict the next most probable word and display it to the right of your screen.

• Sometimes, the model can not predict the next word due to the word you are typing is pretty strange for the model. The app will throw a special token <unknown> to say that it can not deal with the input sentence.

• The link to this app is provided in the next slide.

• Shiny app: https://ztp34z-th0h0i-nguy0n.shinyapps.io/swiftkey.

• Source code: https://github.com/nthehai01/swiftkey.