基于遗传算法和神经网络的软件界面美感建模  

A new software interface aesthetics model based on genetic algorithm and neutral network

作　　者：袁培飒 陈锐[1] 林涛[1] 李枭[1] 

机构地区：[1]四川大学计算机学院,成都610065

出　　处：《四川大学学报（自然科学版）》2015年第2期269-274,共6页Journal of Sichuan University(Natural Science Edition)

基　　金：四川省科技支撑项目(2013GZX0138;2012GZ0091)

摘　　要：用户界面作为人与机器之间信息交互的桥梁,良好的用户界面能够提高人机交互的效率.本文以软件界面布局和色彩特征作为研究对象,采用神经网络方法建立特征与软件美感之间的关系模型,遗传算法进行特征选择,找出对界面美感影响较大的特征,并且通过主成分分析,最终提炼出界面美感的设计要素.A well-designed software interface can increasingly improve the efficiency of the human and computer interaction（HCI）process.This paper focuses on the layout and color characteristics of interface.In order to discover the features which have the most influence on interface aesthetic feeling,we adopt machine learning method to build the relationship model between those characteristics and aesthetics of software interface,and then use the genetic algorithm（GA）to extract the important characteristics.At last,the aesthetic design elements of software interface can be refined by the method of the kernel principal component analysis（KPCA）applied to get the optimal feature subset.

关 键 词：遗传算法 神经网络 界面布局 色彩 用户界面 

分 类 号：TP316[自动化与计算机技术—计算机软件与理论]