Fuzz-ART Neural Networks for Predicting Chi-Chi Earthquake Induced Liquefaction in Yuan-Lin Area

模糊自適應共振類神經網路預估集集大地震中員林地區之液化潛能

陳俶季(S. G. Chern) ； 胡榮豐(R. F. Hu) ； 張延任(Y. J. Chang) ； 蔡宜峰(I. F. Tsai)

《Journal of Marine Science and Technology》 10卷1期 (2002/06) Pp. 21-32

https://doi.org/10.6119/JMST.200206_10(1).0004

摘要

本文以自適應共振理論（ART）為構思基礎，結合模糊系統理論，提出模糊推理類神經網路“Fuzz-ART”來評估集集大地震中員林地區的液化潛能。本研究模式以倒傳遞演算法作為參數學習機制及以模糊自適應共振理論（fuzzy ART）作為架構學習機制。經由案例分析結果顯示，“Fuzz-ART”預估液化潛能可得到較傳統類神經網路更為準確有效之評估結果。若能收集更多關於集集地震的資料，此架構對員林地區之液化潛能將能得到更為良好之評估效果。

並列摘要

In this study, a fuzzy adaptive network “Fuzz-ART”, based on adaptive resonance theory (ART) combined with fuzzy set theory is developed to evaluate liquefaction potentials induced by Chi-Chi earthquake in Yuan-Lin area. The proposed system combines the backpropagation algorithm for parameter learning and the fuzzy ART algorithm for structure learning. With the help of case studies, it is shown that the “Fuzz-ART” network is able to predict liquefaction potentials much more satisfactorily than conventional artificial neural network methods. If more data are collected, it may well evaluate liquefaction potentials induced by Chi-Chi earthquake.

並列關鍵字

Fuzz-ART ； ART ； fuzzy rule ； geotechnical engineering