This Virtual Nietzsche is a predictive sentence generator in the style of Friedrich Nietzsche. We have built it from scratch in order to better understand n-grams.
In addition, a classifier is also trained to identify from which of the five books the line is most likely to be found in.
We have used a dataset from Kaggle (https://www.kaggle.com/hsankesara/books-of-friedrich-nietzsche). This dataset contains five books written by Nietzsche in the form of a text file.
The contents of the books were extracted and a new file (clean_data.txt) has been generated to compile all the sentences (with [BEGIN] and [END] tags generated to signify the beginning or end of a sentence).
Unigrams: P(W_i)
Bigrams: P(W_i | W_(i-1))
Trigrams: P(W_i | W_(i-1), W_(i-2))