Using WordNet-based semantic similarity measurement in External Plagiarism Detection Notebook for PAN at CLEF 2011
Yurii Palkovskii, Alexei Belov, Iryna Muzyka · 2011
Continuing our previous work started at PAN 2009 and PAN 2010 [7] we considered further research options based on the achieved baseline of the best performing algorithms. The research done by Potthast et al. [4] presented a sliced view of the presented approaches showing their performance on specific corpus metrics external\intrinsic, obfuscation strategies (none, artificial high\low, simulated, translated), topic match, case length and document length thus defining the baseline for further studies. A brief analysis of the above named results [1,3] shows that there exists a direct correlation between the obfuscation degree (method) and the achieved performance.