Fish habitat evaluation systems considering spatiotemporal dynamics

本计画旨在发展一个能处理鱼类栖地偏好时空动态的栖地评估系统。两个混合模型的遗传Takagi-sugeno模糊模型和模糊神经网络模型的开发和评价其精度和适用性在生态学。本研究主要取得了以下四个方面的成果：（1）日本青鳉栖息地偏好日变化的实验室研究;（2）遗传Takagi-sugeno模糊模型优化的新方法;（3）数据质量对栖息地偏好模型影响的评价;（4）模糊栖息地偏好模型在顶口黑鲷中的应用。

This project aimed to develop a habitat evaluation system that can deal with spatiotemporal dynamics of fish habitat preference. Two hybrid models of a genetic Takagi-sugeno fuzzy model and a fuzzy neural networks model were developed and evaluated for its accuracy and applicability in ecology. Four major achievements in this project are as follows :(1) laboratory experiment on the diurnal changes of habitat preference of Japanese medaka,(2) new method for optimization of a genetic Takagi-sugeno fuzzy model,(3) evaluation of effects of data quality on a habitat preference model,(4) application of fuzzy habitat preference model to topmouth gudgeon.