AHP与机器学习的矢量数据质量评估

Quality evaluation of vector data based on AHP and machine learning

为了提升新型基础测绘矢量数据成果质量评价效果,提出了基于层次分析法和机器学习的新型基础测绘矢量数据成果质量评价方法。以实用性、全面性、可操作性与层次性为原则,建立新型基础测绘矢量数据成果质量评价指标体系;利用三标度法与云模型改进层次分析法,通过改进层次分析法计算评价指标权重;在模糊神经网络内,输入评价指标与对应的权重,输出新型基础测绘矢量数据成果质量评价结果。实验证明:该方法构建的指标体系内指标相关性较小,即指标体系涵盖信息的全面性较优,信息不重叠程度较低;该方法可有效评价新型基础测绘矢量数据成果质量。

In order to improve the quality evaluation effect of new basic surveying and mapping vector data results,a new basic surveying and mapping vector data results quality evaluation method based on AHP and machine learning was proposed.Based on the principle of practicability,comprehensiveness,operability and hierarchy,a new quality evaluation index system of basic surveying and mapping vector data results was established.Three scale method and cloud model were applied to improve the analytic hierarchy process,and the weights of evaluation indexes were calculated by the improved analytic hierarchy process.In the fuzzy neural network,the evaluation index and the corresponding weight were input,and the quality evaluation result of the new type of basic mapping vector data was output.Experimental results showed that the correlation of indicators in the index system constructed by this method was small,that is,the index system covered better information comprehensiveness,and the degree of information non-overlap was low.This method could effectively evaluate the quality of new basic mapping vector data.