Functional search in economics using genetic programming |
| |
Authors: | Carl P Schmertmann |
| |
Affiliation: | (1) Department of Economics, Florida State University, 32306-2045 Tallahassee, FL, USA |
| |
Abstract: | This paper discusses economic applications of a recently developed artificial intelligence technique-Koza's genetic programming (GP). GP is an evolutionary search method related to genetic algorithms. In GP, populations of potential solutions consist of executable computer algorithms, rather than coded strings. The paper provides an overview of how GP works, and illustrates with two applications: solving for the policy function in a simple optimal growth model, and estimating an unusual regression function. Results suggest that the GP search method can be an interesting and effective tool for economists. |
| |
Keywords: | genetic programming genetic algorithms evolutionary search optimal growth econometrics nonparametric regression |
本文献已被 SpringerLink 等数据库收录! |
|