基于BP神经网络的船员适任性评价模型  

Model of crew competency evaluation based on BP neural network

作　　者：胡志武[1] 程葆明[1] 陈延才[1] 

机构地区：[1]上海海事大学商船学院,上海201306

出　　处：《上海海事大学学报》2010年第4期23-27,共5页Journal of Shanghai Maritime University

基　　金：上海市自然科学基金(09ZR1421500)

摘　　要：为弥补模糊综合评判等方法在船员适任性评价上的不足,采用BP神经网络的学习算法,建立船员适任性评价系统模型.通过分析影响船员适任性的主要因素,结合专家经验和统计数据,确定神经网络的神经元个数,取得训练样本,建立评价指标体系.借助于MATLAB,用某航运公司55份船员数据对神经网络进行训练得到的误差曲线表明,该方法对船员适任性可作出有效评价,具有实用价值.In order to compensate the inadequacy of fuzzy comprehensive evaluation in crew competency evaluation,the learning algorithm of BP neural network is adopted to establish the evaluation model of crew competency.By analyzing the main factors which may influence crew competency,and using the experts’experience and the statistics,the number of neurons is determined,training samples are obtained,and the evaluation index system is established.The error curve obtained from neural network training by using the data of 55 seafarers of a shipping company with the help of MATLAB shows that this model can evaluate crew competency effectively and has practical value.

关 键 词：船员适任性 评价模型 BP神经网络 MATLAB 

分 类 号：U676.2[交通运输工程—船舶及航道工程] TP183[交通运输工程—船舶与海洋工程]