Figure 2

Comparison between the disambiguation performed by different WSD algorithms and the ROUGE-2 scores obtained by the summaries generated using each WSD algorithm. Disambiguation performance is represented in terms of the proportion of common mappings between each pair of WSD algorithms, while their performance in the summarization task is pictured in terms of the difference in the ROUGE-2 scores achieved by the summaries generated. Each data point in these graphs represent a different document from the evaluation corpus. MRD stands for Machine Readable Dictionary, AEC stands for Automatic Extracted Corpus and JDI stands for Journal Descriptor Indexing.