This paper discusses the concept of an IQM (Informal Qualitative Model) which is seen as a bridge between rigorous, and often intractable, theories on the one hand, and experimental data on the other. We argue that the selection of variables to be explored using quantitative law discovery should be made using background knowledge. However, domain theories are often intractable, and to make progress it is therefore necessary to add assumptions; i.e. one is forced to take particular and often simplifying, perspectives on the domain. IQMs essentially capture these ideas. This paper demonstrates how a set of IQMs for a domain (colligative properties of solutions) can generated from a base IQM and a set of operators.
|Title of host publication||Proceedings of AAAI 1995 Spring Symposium on Systematic Methods of Scientific Discovery|
|Number of pages||5|
|Publication status||Published - 1995|
- scientific discovery
- computational model