Metta Search Engine for Medical Literature

Research on a text mining pipeline which aims at accelerating systematic reviews in evidence-based medicine and developed a meta-search engine for medical literature retrieval which currently incorporates five medical search engines: PubMed, EMBASE, CINAHL, PsycINFO and Cochrane Central Register for the first phase of the pipeline. The future related research problems include: (1) creating a supervised learning based literature classification and a learning-to-rank system that take as input the list of retrieved articles by the meta-search engine with respect to a given query, classify them by article type, and rank them in order of predicted probability of relevance to an individual writing a systematic review on the topic; (2) investigating how to create a study aggregator that collects together articles that refer to the same underlying clinical trial, which will provide automated assistance to reviewers in determining whether two articles derive their evidence from the same primary data collection activity. The meta-search engine is available for demo: http://mengs1.cs.binghamton.edu/MetaSearchEngine/

The basic idea is based on the following paper.