Association rule learning is a rule-based machine learning method for discovering interesting relations between variables in large databases.1

It is intended to identify strong rules using measures of interestingness, and the best-known constraints are minimum thresholds on support and confidence.1 Course material describes the method as first used for market basket analysis, to find how items purchased by customers are related, and says that confidence measures the reliability of the inference made by a rule.2

In SWEET the class is a subclass of SWEET’s parameterisation class within the data service analysis vocabulary.3 The retained sources do not establish any Earth observation application, since both descriptions of the method come from general data mining. They also do not establish how the SWEET class relates to the Earth observation processing chain beyond its place in that vocabulary.

Footnotes

  1. Wikipedia, “Association rule learning”. https://en.wikipedia.org/wiki/Association_rules ↩ ↩2

  2. University of Ioannina, “Chapter 6: Association Rules (data mining course slides)”. https://www.cs.uoi.gr/~arly/courses/dm/slides/7.association-rules.pdf ↩

  3. ESIP Federation (SWEET), “SWEET reprDataServiceAnalysis.ttl”. https://raw.githubusercontent.com/ESIPFed/sweet/master/src/reprDataServiceAnalysis.ttl ↩