Knowledge Graph
This map connects the concepts in our Glossary with the articles and topic categories of The Health Topics. Every circle is a concept or a category; every line is a relationship found in the articles themselves. Use it to see how ideas such as patient safety, measurement and artificial intelligence meet, and to jump straight to the articles that discuss them.
Method and limitations
Data
The map is built from the published articles on The Health Topics, the topic categories they are filed under, and the terms defined in the site Glossary. Only glossary terms that occur in at least one article are shown ( concepts). The index is rebuilt when articles or glossary entries change; the date of the current build is shown above.
Concept detection
Each article is searched automatically for glossary terms. Reference lists are excluded. A concept is counted once per article, however often it is mentioned. Two ambiguous terms, reliability and validity, are counted only when they occur in a measurement context.
Links
A solid line joins a concept to the category in which most of its articles are filed, and to a second category when at least two of its articles are filed there. A dashed line joins two concepts that occur together in at least three articles. Co-occurrence links are added from the strongest pair downwards, and a pair is skipped once both concepts already have three links, so the map shows each concept's strongest associations and not every co-occurrence. Line width reflects the number of shared articles.
Layout and colour
Categories are placed on a ring in alphabetical order. Concepts are positioned by a force-directed layout that pulls each one towards its main category. Node size shows the number of articles that mention the concept. Colour shows the main category and is an editorial classification, not a cluster detected from the network.
Limitations
- The map describes what has been published on this site. It is not a map of the wider scientific literature.
- Detection relies on string matching, so a term can be counted where it is used in another sense, and concepts expressed in other words are missed.
- Link weights are raw co-occurrence counts and are not normalised, which favours concepts that appear in many articles.
- Distances on the map have no metric meaning, and the layout can differ slightly between visits.