Introduction

the presentation of the final project from Coursera Data Science Course, this project consist on create a prediction algorithm, and to provide an interface that can be accessed by others.

We have a shiny app that takes a input phrare (one or multiple words) in a text box input and outputs a prediction of the next word.

• Shiny app here

Word Prediction Algorithm

1. Algoithm checks how many words was imput.
2. Select the n-grams table.
3. The most frequent next word will be shown as the return value.
4. If not find the word in the n-gram table, The output will be “The” (The most commond word).
   Code is available on GitHub Further work can include expanding both the number of data sources & number of n-grams Enjoy the app!

Data Gathering & Cleansing

• Merged data from the 3 Data Sources into one data file (Blogs, Twitter & News).
• Cleansed data including converting to lowercase, stripping white space, and removing punctuation & numbers.
• Created Bigram, Trigram and Quadgram n-grams.
• Extracted term-count tables from the n-grams.
• Sorted in descending order based on frequency.
• Saved n-gram objects.

Experience using the app

The final data product is created as shiny app and is available here

An input field takes a text, single word or sentence.
The result (next word) will be displayed on the right side.
In addition, the dataset which is used for the prediction (bigram, trigram, qudgram) is displayed.
Application is available on , see link above.

1. The app uses the correct prediction algorithm for words prediction.
2. The results shown uses the functions for prediction.