The lib for quantitative analysis of textual data - quanteda
Kneser-Ney smoothing based on probabilistic language model
Prediction algorithm allows to get the third word max
Result is presented as a web - app
Kneser-Ney algorithm
Making N-gram model as probability distribution over strings that attemps to reflect how frequently a string occurs as a sentence based on the counts in a traning set.
Applying smoothing - calculating probabilities by taking into consideration order of words and their context: the unigram probability used should not be proportional to the number of occurrences of a word , but instead of to the number of different words that it follows.
Put one or two word into the text field for prediction
Expected result:
The app would show five the most probabl word if there is prediction based on the your input, otherwise shown the most probable words computed regardless of input.
