SFX article recommendation service
- What is the recommendation service?
- How do I use the recommendation service?
- Why are there sometimes no recommendations?
- How are the recommendations derived?
- Sample searches
When you are viewing the SFX link to a particular article, the service recommends other articles which may be similar to the one you've requested. These recommendations are based on real usage data from thousands of researchers across the world, and are updated continually.
This type of service may be familiar to users of services such as Amazon, which - based on a purchase that you have in common with other customers - present suggestions for other items that might interest you.
If you are using the SFX menu to view an article, you may see a box of up to 3 recommendations appearing in the menu. Please note that if no recommendations are available you will not see the box. An example is shown below.
You then have various options:
- If the title of a recommended article is underlined, click on this title link to go straight to the full text of the article
- Click on the SFX button at the right of any article to see an SFX menu for the new article, including the option to search library catalogues, and new recommendations based on that article.
- Click on "View More" to see up to 10 recommendations
- Select one or more (or all) of the recommendations and select a format in which to export these, from the list next to "Save Citations". This list includes bibliographic software, such as Reference Manager or EndNote.
According to analysis by the service provider, users should see recommendations in approximately 40% of cases of the SFX menu. There are several reasons why you might not see recommendations:
- Recommendations are only provided for journal articles, not for whole journals, nor for books. So, if you have reached the SFX menu through UCL's list of electronic journals, for example, no recommendations will appear.
- Often, newly published articles have not yet been viewed enough to generateassociations with other articles.
- Some very specialist articles may only have been viewed a handful of times and therefore do not have sufficient user data to produce meaningful recommendations.
The service analyses anonymised usage data from institutions that use SFX, worldwide. It uses algorithms based on scholarly research to ascertain the strength of links between articles, based on the real usage data. The stronger the link between two articles, the higher they are up the list of recommendations for one another.
Try searching for the following in a relevant database for a good chance of seeing recommendations:
- influenza vaccination
- urban transport
- stem cell AND cloning
- honeybees (in Engineering Village)
- biomass to liquid
- Hungerford Bridge (in ICONDA)
- mixed sex (in British Education Index and ERIC)
- European Union (in MLA International Bibliography)
To submit feedback or request help with this service, please use the E-resources Contact us form.