基于自适应小批量梯度学习的网络地理信息服务预测系统研究

An Adaptive Mini-Batch Gradient Learning-Based Geographic Information Service Prediction Systems

文章面向REST类型的网络地理信息服务,实现了网络地理信息服务的预测与分析系统。首先,采用爬虫与HTML分析技术,获取了 ESRI网站的共享地理信息服务数据;然后,在完全批量梯度与随机梯度计算模型的基础上,提出了基于自适应小批量梯度学习的网络地理信息服务质量预测模型,结合JQuery前端以及基于ArcGIS REST服务API,实现了网络地理信息服务信息查询、地图服务质量描述与评估、动态地理信息服务性能在线预测等功能的原型系统,更好地为地理信息服务用户发现、选择、使用网络地理信息资源服务,从而使地理信息用户体验到更好的网络地理信息服务。

The key parts of our geo-system are the prediction and analysis function of geographic information service. The Geo-Service data is obtained by searching the Rest service information from the ESRI website. The font web end of our Geo-system is implemented by jQuery. Our Geo-system consists of adaptive mini-batch gradient prediction, and developed quality evaluation, search function and registration and login function, using ArcGIS REST API. The quality evaluation function is developed by multi-objective model to improve the quality of Geo-system services. And the prediction function of the Geo-system can help us predict the quality of Geosystem rest services.