Textsummarization has become an important research area, especially in the biomedical domain, where information overload is a major problem. In this paper, we propose a novel biomedical text summarization system that combines two popular data mining techniques: clustering and frequent itemset mining. Biomedical paper is expressed as a set of biomedical concepts using the UMLS metathesaurus. The K-means algorithm is used to cluster similar sentences. Then, the Apriori algorithm is applied to discover the frequent itemsets among the clustered sentences. Finally, the salient sentences from each cluster are selected to build the summary using the discovered frequent itemsets. For the evaluation step, we selected randomly 100 biomedical papers from the BioMed Central database full-text, and we evaluated the performances of our system by comparing the resulting summaries with the abstracts of these papers using the ROUGE metrics in term of recall, precision, and F-measure. We also compared the obtained summaries with those achieved by five well-known summarizers: TextRank, TextTeaser, SweSum, ItemSet Based Summarizer, Microsoft AutoSummarize, and two baselines: summarization using only the frequent itemsets mining (FRQ-CL), and summarization using only the clustering (CL-FRQ). The results demonstrate that this combination can successfully enhance the summarization performances, and the proposedsystem outperforms other tested summarizers.

Introduction

The development of the World Wide Web, especially in the last two decades has led to an exponential growth of online information. This is also the case in the biomedical domain, e.g., MEDLINE1 (med), the largest biomedical bibliographic text database contains about 25 million references of journal articles in life sciences that concentrate on biomedicine. However, researchers in this area encountered major difficulties to access to the desired information quickly and efficiently (Afantenos, Karkaletsis & Stamatopoulos, 2005). Text summarization is a promising technique that could aid them to obtain the core information in a given subject by “condensing the source text with preserving the main ideas from it” (Mishra, Bian, Fiszman, Weir, Jonnalagadda, Mostafa & Del Fiol, 2014). i.e., text summarization could aid biologists to find general information about a biological concept, e.g., a gene or a disease, from one or multiple documents without reading the entire documents (Shang, Li, Lin & Yang, 2011). Medical doctors frequently use summaries to identify patient’s treatments quickly, and to reducing diagnosis time (Reeve, Han, Nagori, Yang, Schwimmer & Brooks, 2006b). Furthermore, summaries are also used to improve indexing and categorization of biomedical papers when it is used as a substitution of abstracts when they are not available (Gay, Kayaalp & Aronson, 2005). The majority of text summarization methods do not consider the characteristics of the domain or the type of documents.