Discovering Admissible Simultaneous Equation Models from Observed Data

Conventional work on scientic discovery such as BACON derives empirical law equations from experimental data. In recent years, SDS introducing mathematical admissibility constraints has been proposed to discover rst principle based law equations, and it has been further extended to discover law equations from passively observed data. Furthermore, SSF has been proposed to discover the structure of a simultaneous equation model representing an objective process through experiments. In this paper, SSF is extended to discover the structure of a simultaneous equation model from passively observed data, and is combined with the extended SDS to discover a quantitative simultaneous equation model re ecting the rst principle.

• Type: Conference
• Year: 2001
• Where: ECML
• Authors: Takashi Washio, Hiroshi Motoda, Yuji Niwa
• Keywords: ECML 2001, Law Equations, Machine Learning, Rst Principle, Simultaneous Equation Model