New PubPharm tool: Recommender for similar articles and clinical trials

PubPharm now offers a feature that suggests similar articles and clinical studies for each result in the detailed view. This recommender tool utilizes the Best Matching 25 (BM25) algorithm to identify related documents.

To process abstracts in PubPharm, “stopwords”, words that do not contribute meaning in context, are removed. Next, the remaining words are reduced to their base forms through a process called “stemming.” This allows calculating document and word frequencies (“term frequency, inverted document frequency”). By comparing references, similar documents can be identified and displayed.