Qualitatively Constrained Equation Discovery
نویسندگان
چکیده
Equation discovery is a very lively area of artificial intelligence which deals with explaining phenomena by mathematical formulae induced from the data. One successful approach to the problem are algorithms which construct thousands of formulae and report the simplest ones with the best fit to the data. Another, sub-symbolic, fits (piecewise) regression hyper-planes; their advantage is that they may be made to conform to qualitative constraints. We propose an algorithm that shares the qualities of the two approaches: EDGAR searches for simple qualitatively faithful equations which fit the data well. The algorithm performs very well on simple problems, but in its current implementation fails to solve more complex ones.
منابع مشابه
An equation-discovery approach to earthquake-ground-motion prediction
In active seismic regions an earthquake’s peak ground acceleration (PGA) is required information when designing a building. In this study we employ the state-of-the-art, Lagramge, equation-discovery system to induce an equation that is suitable for modeling the PGA and investigate its applicability. In contrast to traditional modeling techniques the Lagramge system does not presume the structur...
متن کاملCausality on cross-sectional data: Stable specification search in constrained structural equation modeling
Causal modeling has long been an attractive topic for many researchers and in recent decades there has seen a surge in theoretical development and discovery algorithms. Generally discovery algorithms can be divided into two approaches: constraint-based and score-based. The constraint-based approach is able to detect common causes of the observed variables but the use of independence tests makes...
متن کاملA matrix LSQR algorithm for solving constrained linear operator equations
In this work, an iterative method based on a matrix form of LSQR algorithm is constructed for solving the linear operator equation $mathcal{A}(X)=B$ and the minimum Frobenius norm residual problem $||mathcal{A}(X)-B||_F$ where $Xin mathcal{S}:={Xin textsf{R}^{ntimes n}~|~X=mathcal{G}(X)}$, $mathcal{F}$ is the linear operator from $textsf{R}^{ntimes n}$ onto $textsf{R}^{rtimes s}$, $ma...
متن کاملLearning Markov Network Structures Constrained by Context-Specific Independences
This work focuses on learning the structure of Markov networks. Markov networks are parametric models for compactly representing complex probability distributions. These models are composed by: a structure and numerical weights. The structure describes independences that hold in the distribution. Depending on the goal of learning intended by the user, structure learning algorithms can be divide...
متن کاملTrivially related lax pairs of the Sawada-Kotera equation
We show that a recently introduced Lax pair of the Sawada-Kotera equation is nota new one but is trivially related to the known old Lax pair. Using the so-called trivialcompositions of the old Lax pairs with a differentially constrained arbitrary operators,we give some examples of trivial Lax pairs of KdV and Sawada-Kotera equations.
متن کامل