Introduction to CBML

Hayes et al. (1999) describe imposing a standard case-based view on the information system of an application in order to retrieve case-knowledge. They anticipated that a standard way of representing case based reasoning (CBR) information will make this easier to achieve and proposed a case representation language that would facilitate this. Without such a standard means of representing case data it is up to the application developer to shape the case data from the available knowledge base. The manipulation of case data is dependent on the representation format chosen by the developer. This limits transformation of the data into a format suitable for the presentation layer, or its movement to another back-end component or between distributed CBR components. Hayes et al. proposed a standard case representation language called Case-Based Mark-up Language (CBML) in 1998. Our work in the field of CBR representation is a continuation of that work. A good description of CBML is given in Chapter 4 of my PhD. thesis.


The original CBML homepage was on Trinity's Machine Learning Group website. There are some archived resources there. The XML Schema file that defines CBML version 3 is here and a sample java file that uses CBML is here (this depends on the original CBML codebase). CBML was the representation format used in the Fionn machine learning framework. If you're interested in more information about CBML, or want access to the source code, contact Lorcan.

CBML Publications