ZB MED: Information Centre for Life Sciences

Intelligent Retrieval – Optimisation of MEDPILOT and GREENPILOT

The subject portals MEDPILOT and GREENPILOT offer up-to-date search technologies coupled with computational linguistic methods. The data from MEDPILOT and GREENPILOT have been merged in a combined index in a joint project with averbis GmbH using the open source Apache Lucene framework.

Morphosaurus

Morphosaurus
Morphosaurus

Searches in MEDPILOT and GREENPILOT have been expanded to include a method of semantic analysis known as Morphosaurus, which has significantly improved the quality of the search results presented to users.

Morphosaurus breaks down individual search terms into their smallest components, known as morphemes. It lists sub-words as word classes that are intelligently linked to specialist thesauri in multiple languages. This makes it possible to search not only for complete words in a certain language, but also for similar words and word associations.

An example of how Morphosaurus breaks down words into smaller components: The German word "Herzmuskelentzündung" (heart muscle inflammation) is segmented into the German morphemes "Herz", "Muskel" and "Entzündung", which are then depicted in relation to the root words "HEART", "MUSCLE" and "INFLAMMATION". This enables the search engine to return translations of the original German word such as "of the heart muscle" as well as specialist terminology such as "myocarditis".

Intuitive search

As well as the basic search engines, the subject portals MEDPILOT and GREENPILOT also include intelligent functions that enable users to perform optimised, intuitive searches.

A database of 25 million entries means that many search terms produce large numbers of relevant hits. To reduce the number of hits, users of traditional subject portals are required to hone their search by entering more precise details in the various different search fields. Although this may make sense for an expert search mode, it does not enable users to quickly and effectively track down the records they need if they are not entirely familiar with exactly how the search engine works.

MEDPILOT and GREENPILOT attempt to remedy this by using the Morphosaurus functionalities of the search engines to provide faceted navigation and intuitively guide users to the results they are looking for. Users are presented with relevant categories in which they can narrow their search such as year of publication, document type, language and database, together with associated subcategories. They can progressively hone their search by clicking on the corresponding categories until the number of hits has become more manageable.

In addition, the system displays associated terms from the respective subject areas that match the search terms entered by the user, allowing the number of hits to be narrowed still further.

ZB MED is working closely with averbis GmbH on an ongoing project to continuously optimise the search engines and their integration in MEDPILOT and GREENPILOT.

Get in touch.

List of contacts