Background: Retrieving gene and disease information from massive biomedical articles to provide doctors with clinical decision support is one of the important research directions of precision medicine .
Methods: we present a new method for biomedical article retrieval based on co-word analysis and cuckoo search. The specific goal is to retrieve biomedical articles, in the form of article abstracts, addressing relevant treatments for a given patient. The method in this paper first uses the BM25 algorithm to calculate the score of the abstract, and we designed a method based on BM25 to calculate the score of expanded words. Second, when a disease and a gene both appear in the same biomedical article, the score of the biomedical article will be increased. Finally, the cuckoo algorithm is used to optimize the parameters of the retrieval algorithm. The paper discusses the influence of different parameters on the retrieval algorithm, and summarizes the parameters to meet different retrieval needs.
Results: all data were taken from medical articles provided in the TREC (Text Retrieval Conference) Clinical Decision Support Track 2017、2018 and 2019 in precision medicine. 120 standard topics were tested. we chose 3 test indicators and many kinds of algorithms for experimental comparison. For the fairness of the experiment, all these selected algorithms all used the BM25 algorithm or an improved BM25 algorithm. Experimental results showed that our algorithm has achieved good results and ranking
Conclusions: we designed an improved BM25 algorithm based on co-word analysis and cuckoo search and verified the superiority of our algorithm on a large number of experimental sets. In this paper, the query expansion method is relatively simple, the next step is to consider the ontology and semantic network to expand the query vocabulary.