An Effective Approach to Estimating Computing Time of Vector Data Spatial Computational Domains in WebGIS
Mingqiang Guo, Ying Huang, Zhong Xie, Liang Wu · GEOMATICA · 2017
Computing time estimation is an arduous issue for scientists in computer science and GIScience. In order to build a more effective estimation model for computing time of spatial computational domains (SCDs) in WebGIS, decision tree machine learning method is leveraged to build a computing time decision tree (CTDT) model. The CTDT modelling approach is focused on and elaborated in this paper. The node splitting method is the key technology of this new approach. It can effectively address the issue of computing time estimation. The computing time estimation framework of SCDs in WebGIS is developed by this study. Since the learning samples of SCDs have been collected, the CTDT model of computing time of SCDs in WebGIS can be easily trained. To demonstrate the effectiveness of the new approach, map visualization is chosen as a typical SCD in WebGIS to conduct a group of experiments. The test results indicate that the performance of CTDT is obvi ously higher than area method (AM) and regression analysis method (RAM). It is capable of estimating the computing time of SCDs. The effective computing time prompt on the client side can tremendously improve the user’s interactive experience.