Image source: biolifejournal
The study by Zhiyong Lu and other collagenous of the National Center for Biotechnology Information (NCBI), published in open access magazine PLOS Biology, shows the world's largest biomedical literature database results by sequentiality, rather than by date to provide a better experience for the users of PubMed.

PubMed contains more than 28 million articles from biomedical literature, in which two more are added in every minute. This is a mandatory resource that is global in scope and is used by millions of users every day. 

From its beginning, search results were returned only in reverse chronological order, most recently, a ranking system emphasizing repetition rather than the relevance of the search query. In 2013, a relevance ranking system was introduced, but it relied on artificial weight lifting factors and requires constant manual adjustment.

In June 2017, the employees of National Center for Biotechnology Information introduced a Machine Learning algorithm which attracts dozens of relevance signals, including user responses, to improve relevant reactions, in particular, click through clicks on given articles for a given search. This ranking system, called "Best Match", is presented as an alternative to chronological ordering.

The team found that click-through rate has increased the results returned by the best match by 20% compared to the results presented in chronological terms. The total utilization of relevancy ordering has increased by 7.5% of all searches before the beginning of the best match till 12% by April 2018. The Machine Learning system relies on user input for improvement.

For more Trending News on Machine Learning, Stay Tuned with us on FoxSplit.