摘要
贝叶斯网络具有强大的推理能力,能与先验知识和数据结合,进行定性和定量分析,提供了1条有效的处理预测问题的途径,本文首先介绍了贝叶斯网络基本理论及其特点,并讨论如何学习贝叶斯网络结构,然后由专家知识和给定数据,采用基于依赖分析的贝叶斯网络学习算法构造了海底网箱养殖水环境指标间的贝叶斯网结构模型。该模型能有效的表达网箱养殖环境各个指标之间的因果关系和影响程度,实验结果表明,试验数据显示准确性是92.3%,kappa指数是0.882。以上证明该方法是有效可行的,表明贝叶斯网络是一种很有前途的预测评价方法。
Bayesian network has a powerful ability for reasoning and semantic representation, which combined with quality and quantitative analysis with prior knowledge and observed data, and provides an effective way to deal with prediction. Firstly, in this paper an over view of Bayesian network and its characteristics is presented, and how to learn a Bayesian network structure from given data is discussed, and then a model of the Bayesian network is constructed by the dependency analysis of Bayesian network learning algorithm for the period of Xiangshan Baycage culture with expert knowledge and the dataset. The model can be effectively expressed in various cage aquaculture environment for a causal relationship and impacts between indicators. The experimental results based on the test dataset are that evaluation accuracy is 92.3 %, and Kappa index is 0. 882. All these prove that the method is feasible and efficient and indicate that Bayesian network is a promising approach for the prediction.
出处
《海洋湖沼通报》
CSCD
北大核心
2009年第1期135-140,共6页
Transactions of Oceanology and Limnology
基金
宁波市海洋渔业局(甬海办2005/331-6)
关键词
象山港
网箱养殖
模型
贝叶斯网
依赖分析
Xiangshan Bay
cage aquaculture
model
Bayesian network
dependency analysis