About the App

This application uses the last English word of your text input. You can predict the upcoming one to four words by selecting the number of words you want to predict. You can access the app here.

Algorithmn for text prediction

• Predictive text uses algorithms trained on large amounts of text data to anticipate your next words.

• The training data is sampled from the English text data provided by SwiftKey.

• Tidy principles of tidyverse and tidytext functions are applied to the data to create N-grams and calculate their Term Frequency-Inverse Document Frequency (TF-IDF) per line.

• N-grams with the most highest TF-IDF scores are used to predict the next words.

How does the app work?

• Uses SwiftKey data from blogs, news, and Twitter to understand common phrases.

• Creates word sequences to predict the most likely next words.

• Identifies the most important words (TF-IDF score) in a sentence for better predictions.

• Put your words or phrases in the input text box, and choose how many words you want to predict.

Output of the App

If you put “happy” in the text input, and select 4, the app will return a list of 4 upcoming words with their score of importance.