Height Conversion in Momentum and Adaptive Learning Rate Algorithm

摘要：

There was a long training time for the norm BP neural network for GPS Height fitting, and easily converging to local minimum problems. Paper, introduced momentum and adaptive learning rate algorithm to improve the norm BP neural network for resolving the problem of the training and convergence, compared with the standard neural network, and calculating by a regional elevation control point coordinates, additional momentum adaptive neural network algorithm accuracy of GPS height conversion was much higher and more stable, and the convergence was much faster..

关键词： GPS additional momentum algorithm adaptive learning rate algorithm height conversion

作者: Hu Chuan

作者单位: Institute of Engineering Surveying Sichuan College of Architectural Technology Deyang, China

会议类型: 国际会议

会议名称: 2010 International Conference on Computer,Mechatronics,Control and Electronic Engineering(2010计算机、机电、控制与电子工程国际会议 CMCE 2010)

会议地点: 长春

会议语种:英文

页码: 92-95

在线出版日期: 2010-08-24（万方平台首次上网日期，不代表论文的发表时间）