Course Data Science Capstone

Word prediction app Presentation

• author: Leandro Freitas
• date: 04/24/2018

The application:

• Takes as input a phrase (multiple words) in a text box;
• Outputs a prediction of the next word on another text box; and
• Works only with text in English.

• First was created a data sample from the imported data available to the project;
• Then this sample was cleaned by conversion to lowercase, removing punctuation, links, whitespace, profanation words, numbers and all kinds of special characters;
• The data sample was also tokenized into bi-,tri- and quadgram;
• The n-gram term frequency matrices have been transferred into frequency dictionaries;
• Resulting data frames are used to predict the next word for the text inputed based on frequencies of the underlying n-grams disctionaries.

To test the app, please visit: https://leandreitas.shinyapps.io/Word_Prediction/.