返回Google圖書搜尋

Reduction Reconsidered: Further Investigation of a Mechanism for Searching for Regularity in Data

Simon Fraser University. Centre for Systems ScienceH. J. Hamilton

出版	Simon Fraser University, Centre for Systems Science, 1989
URL	http://books.google.com.hk/books?id=xmL8ewEACAAJ&hl=&source=gbs_api

註釋Recently, Wu (Wu 88) proposed a method called Reduction for searching for regularities in experimental data. Reduction was described as a polynomial-time algorithm that is both more efficient than the methods used in BACON, ABACUS, and COPER, and sufficiently powerful to make significant scientific discoveries in uncultivated domains, meaning a scientific field where observational data are being accumulated and need to be analyzed for numerical relationships. This paper examines Reduction to identify the conditions required to permit its claimed efficiency and the corresponding limitations in its applicability. Reduction is briefly described and an algorithm is presented based on the description of the method given in Wu's paper. Also, complexity analyses of Basic and Universal Reduction are given. Unstated assumptions made in the design of Reduction are explicitly documented. The particular class of problems for which Basic Reduction is applicable is characterized, and a more efficient algorithm for such problems is proposed. Conclusions and suggestions for further research are also presented.