Intrinsic features of biomedicai document for the efficient single document summarization

In this paper we demonstrate if the similarities between the abstract and individual sections comprised of the document are increased when we apply our devised additional weights based on intrinsic document features for better effective abstraction of a single scientific article belong to biomedical domain. And we treat MLP based experimental results about the prediction of presence or absence of the terms in the abstract, which are with additional weights appeared in the rest of paper except the abstract.