Pr3d1ct0 is an R-based ShinyApp created in the context of the Coursera Data Science Capstone Project.
Aim: Explore different text prediction algorithms (n-grams linear interpolation, Stupid back-off and Katz back-off).
7/1/2020
Pr3d1ct0 is an R-based ShinyApp created in the context of the Coursera Data Science Capstone Project.
Aim: Explore different text prediction algorithms (n-grams linear interpolation, Stupid back-off and Katz back-off).
| Number of ngrams | Model size | |
|---|---|---|
| 1-gram model | 8,155 | 0.7 Mb |
| 2-gram model | 143,366 | 16 Mb |
| 3-gram model | 76,392 | 11.1 Mb |
| 4-gram model | 99,872 | 19.1 Mb |
| 5-gram model | 33,756 | 8.3 Mb |
Visualization: predictions can be displayed as bar chart or “cloud of words”.
Automatic prediction when SPACEBAR is pressed!