应用ANN与CA构建GIS模型的城市非正规商业行为分析

基于神经网络算法(ANN)和元细胞自动机(CA),建构台湾逢甲夜市商圈的地理信息系统(GIS)模型,并对非正规摊贩商业行为进行数值模拟与实证分析.研究结果表明:非正规摊贩分布模拟结果为平均正确率84.1%,而目标模式的平均正确率可达94.2%;在环境因子影响性分析方面,街道位置类型是发生非正规摊贩最具影响性的环境因子,而背景建物宽度则是最不具影响性的环境因子;在环境因子相关性分析方面,街道位置类型因子是所有假设环境因子中最能吸引非正规摊贩聚集的正相关因子.

Analysis of Urban Informal Business Behavior Based onGIS Model Constructed by ANN and CA

Based on the artificial neural network(ANN)algorithm and the cellular automation(CA), the geographic information system(GIS)model of the Fengjia night market in Taiwan was constructed, the commercial behavior of the informal street vendors was simulated. The results show that: the average simulation accuracy of the informal street vendor distribution is 84.1%, but the average simulation accuracy of target pattern is 94.2% high. In the environmental factor impact analysis, the street location type is the most influential environmental factor of the informal vendors. The background building width is the least influential environmental factor. In the environmental factor correlation analysis, the street position type factor is the positive correlation factor that is the most attractive environmental factor for attracting informal vendors.