作者: Florina Piroi , Aldo Lipani , Mihai Lupu , Allan Hanbury
DOI: 10.1109/ICDAR.2015.7333830
关键词:
摘要: Creating systematic reviews is a painstaking task undertaken especially in domains where experimental results are the primary method to knowledge creation. For review authors, analysing documents extract relevant data demanding activity. To support creation of reviews, we have created DASyR—a semi-automatic document analysis system. DASyR our solution annotating published papers for purpose ontology population. dictionaries not existing or inadequate, relies on annotation bootstrapping based positional Random Indexing, followed by traditional Machine Learning algorithms extend set. We provide an example application subdomain Computer Science, Information Retrieval evaluation domain. The reliance this domain large scale studies makes it perfect test on. show utility through different parameter values bootstrap procedure, evaluated terms annotator agreement, error rate, precision and recall.