Predict the next word in the sentence

Typing on mobile devices can be a serious pain.

I have created a model that will reduce this pain by automatically predicting the next word in the sentence

For the purpose of this project

• I have used data set from twitter, blogs and news
• Explored the data set
• Cleaned the data set
• Built a prediction model
• Deployed as a Shiny App

The raw data set contained a lot of unnecessary characters. I have loaded some of those random lines to explore the data set and clean it

• Removed all the weird characters and numbers
• Transformed the text to lower case
• Split the sentences into individual character elements
• Removed punctuations apart from - and '
• Removed unnecesarry words

• I have used 70% of my data set to build my model.
• I need to build an array that displays the frequency of words in that data set
• I also had to build a frequency of words for 2-grams, 3-grams and 4-grams
• Sort all the n-grams in decending order of probability
• I then storeD our n-grams to the disk to save on memory

• I have loaded all the n-grams from disk for our model
• Get the input sentence from the user
• Select last few words(max 3) from the sentence
• Find those words in our n-grams (4,3,2,1)
• Display the next word from the sentence
• Word Cloud
  

• This application was built as a Shiny App
• Here is the link to the Application –> http://vamshideep.shinyapps.io/predictnextword
• You can enter your words of the sentence and see the predicted output.

Thanks for checking my prediction app