Simplipedia-Making Wikipedia accessible

Abstract: Wikipedia is a massive encyclopedia of information available on the internet, but everyone may not be able to take part of this information. Due to cognitive impairments and other disabilities people may nd it hard to read and understand normal text. A possible solution to making the text more accessible for disabled reader is to augment the text with Blissymbolics, a graphical language made of pictograms. The subject of this thesis is to implement a tool for automatically connect Blissymbolics to the data from Wikipedia articles. The soft- ware generates an XML output from which it is possible to create di erent visualisations which can target di erent groups. The software uses CCF Symbol Server a service which looks up Blissymbolics for words, these symbols is stored in the XML together with the word which was used to look it up. The software also uses a NLP tool from Stanford, from which certain information is extracted and stored in the XML. An evaluation of how well the symbols matches the word was also made, the result shows that it has potential to be very e ective at this, but a symbol selector needs to be implemented. Without a symbol selector it is not as e ective and up to one third of the words may get a bad matching symbol. This report is nally concluded with a discussion and a few ideas of how the improve the result further.