首页 > 电子期刊 > J > 计算机时代

一种自适应停止的SOM神经网络及其应用

rameters need to be set when initializing the network. A self-adaptive method is introduced for determining the network training times, which uses weights derivative to determine whether to stop training. After the test and comparison of the traditional and the improved network by taking advantage of UCI datasets, it turns out that the improved network is better than the traditional network. Finally, when the improved network is used in identification of the vehicle, an expected recognition results can be reached.Key words: SOM neural network; network training; a self-adaptive method; cluster0 引言RFID技术作为一种新兴的车辆监控技术,正在被广泛地应用于交通领域,重庆市目前正在建设全球最大规模的以UHF RFID技术为基础的车联网,已建设RFID数据采集点300余处,给利用RFID数据来进行非法营运车辆的识别提供了可能。项目前期已经得到了模拟城市某点的RFID对各种机动车辆仿真的数据,其中包括营运车辆和非营运车辆。营运车辆包括:公交车、长短途客运车辆和出租车;非营运车辆是私家车。识别的任务就是找到非法的营运车
<<上一页  下一页>>

首页 > 电子期刊 > J > 计算机时代

广州市越秀区图书馆版权所有。
联系电话:020-87673002

本站访问人数: