Parameter Screening and Optimisation for ILP Using Designed Experiments
Reports of experiments conducted with an Inductive Logic Programming system rarely describe how specific values of parameters of the system are arrived at when constructing models. Usually, no attempt is made to identify sensitive parameters, and those that are used are often given “factory-supplied” default values, or values obtained from some non-systematic exploratory analysis. The immediate consequence of this is, of course, that it is not clear if better models could have been obtained if some form of parameter selection and optimisation had been performed. Questions follow inevitably on the experiments themselves: specifically, are all algorithms being treated fairly, and is the exploratory phase sufficiently well-defined to allow the experiments to be replicated? In this paper, we investigate the use of parameter selection and optimisation techniques grouped under the study of experimental design. Screening and “response surface” methods determine, in turn, sensitive parameters and good values for these parameters. This combined use of parameter selection and response surface-driven optimisation has a long history of application in industrial engineering, and its role in ILP is investigated using two well-known benchmarks. The results suggest that computational overheads from this preliminary phase are not substantial, and that much can be gained, both on improving system performance and on enabling controlled experimentation, by adopting well-established procedures such as the ones proposed here.
KeywordsResponse Surface Parameter Selection Full Factorial Design Inductive Logic Programming Fractional Design
Unable to display preview. Download preview PDF.
- 5.Breiman, L., Friedman, J.H., Olshen, R.A., Stone, C.J.: Classification and Regression Trees. Wadsworth, Belmont (1984)Google Scholar
- 6.Dzeroski, S.: Relational Data Mining Applications: An Overview. In: Dzeroski, S., Lavrac, N. (eds.) Relational Data Mining, pp. 339–360. Springer, Berlin (2001)Google Scholar
- 7.King, R.D., Muggleton, S.H., Srinivasan, A., Sternberg, M.J.E.: Structure-activity relationships derived by machine learning: The use of atoms and their bond connectivities to predict mutagenicity by inductive logic programming. Proc. of the National Academy of Sciences 93, 438–442 (1996)CrossRefGoogle Scholar
- 10.Srinivasan, A.: Four Suggestions and a Rule Concerning the Application of ILP. In: Lavrac, N., Dzeroski, S. (eds.) Relational Data Mining, pp. 365–374. Springer, Berlin (2001)Google Scholar
- 11.Steppan, D.D., Werner, J., Yeater, R.P.: Essential Regression and Experimental Design for Chemists and Engineers (1998), http://www.jowerner.homepage.t-online.de/download.htm