engleski

Keyword tagging, extraction and visualisation from multilingual texts

Application described in this thesis is a web application which performs keyword extraction and visualisation on texts written in Croatian, English and Italian language. Keyword extraction is a process of determining which terms best describe a given document with a caveat that such terms must be present in the document itself. Manual keyword extraction is a very time consuming and tedious task making it a perfect candidate for automation. Application is written in Python on top of Django framework and uses Maui topic indexing algorithm developed at the University of Waikato. The thesis serves as a documentation of practical use of the algorithm focused on keyword extraction task without the use of controlled vocabulary and as a documentation of the application architecture and design decisions behind it.

keyword extraction, Maui, web application

nije evidentirano