Centre d’innovació i Tecnologia BARCELONATECH

APPLICATION TO DETECT WRITING PATTERNS

inLab FIB UPC has developed the prototype of an application that can detect writing patterns through natural language processing (NLP) techniques. Using documents written by one person, the application infers a set of characteristic indicators of the way that it is written (words that are used, length of sentences, richness of the vocabulary, and the frequency of use of some words, among other factors). From the patterns that are obtained, the application can determine whether a person was or was not the author of a specific document, and detect the probability of them being the author. Using some classification algorithms (Support Vector Machine and Knn, among others) plagiarism can be detected. This project was developed for the Open University of Catalonia (UOC).