Copy the training corpus (the documents against which plagiarism will be checked) in
The main.py folder and run it.

Copy the content of the file into the text box and give the file a name as it will now be
added in the training corpus. Press the preprocess button to initialize the trigrams
and now we are ready to go.

Press the check button to generate a graph between the percentage plagiarism wrt to
each file in the training corpus.
Now close the graph, take another file if we want to proceed without adding previous file
In the corpus just click the check button.

In case the sentences are jumbled the tri-gram methods still detects it. If the document is
completely copied our model performed very well along with in the case where the text
is unique. The grey areas were the places where the document has been paraphrased
heavily and since the model works on a tri-gram comparison if the words are replaced
with synonyms the efficiency will decrease. This can be overcomed by using a model
Which detects the context of the sentences.