The main objective of this project is to create a shiny application. This shiny application must predict the next word of a phrase, introduced by the user as a text input. First of all, an exploratory data analysis was performed over the data, cleaning it.
Alhorithm
In order to predict the word, different ngrams have been created. Each of the n-gram is used to store the memory of a previously analysed tweets, blogs and news documents.
Unigram (n = 1)
Bigram (n = 2)
Trigram (n = 3)
Quagram (n = 4)
5-gram (n = 5)
Sixgram (n = 6)
How the shiny app words:
The application consists of 2 driving files:
ui.R : User Interface
server.R: back end of system
There are also some files, called ngram_X.rds, that are those where the training data have been saved. The siny application requires the user to introduce an input text in a box. Then, it will predict the 5 most probable words. To find this, it will search for the most common combination of the words entered.
