【摘要】 地理信息系统(GIS)中的属性数据绝大多数是描述性的定性数据,而不是定量数据,有些学者试图应用粗集、云理论、遗传算法和灰色理论等讨论属性数据的精度度量和质量控制方法,但这些方法尚难以在实际工作中应用。据此主要探讨了以抽样检验的缺陷率作为属性数据的精度度量,通过抽样检验获得缺陷率的估计值,采用缺陷率对GIS属性数据进行质量控制的方法。缺陷率是否可靠很大程度取决于抽样方案的选择,依据“限定最大相对偏差”的原则来确定一般随机抽样的抽样样本容量,给出了具体样本容量的计算公式和不同复杂程度的属性数据在不同置信水平下的样本容量,并制成表格以供查表使用 在此基础上探讨并给出属性数据缺陷率度量的质量控制的限差,当属性数据质量的缺陷率不超过该限差则认为属性数据质量是合格的。通过实例说明这种基于缺陷率的质量控制方法可以用于实践,可有效地对属性数据进行质量控制。更多还原

【Abstract】 Most of attribute data in geographic information system (GIS) are described with non - quantity feature data other than quantity feature data. Some scholars discuss the accuracy measurement and the quality control of attribute data with the theory of rough set, fuzzy set, genetic algorithm, gray theory, Chaos theory and so on. However,such methods are still difficult to be used in practice application. In this paper, the rate of disfigurement is discussed to be used to measure the accuracy of attribute data based on the sampling inspection, which is a new method during the researching of accuracy of attribute data in GIS. However, the selection of the sampling schemes can decide the reliability of the rate of disfigurement. So the inspection size of simply random sampling is given according to the principle of limiting the largest relative bias, while the computation equation and result are given and tabled so that they can be consulted. Based on all the above, the tolerance for quality of the attribute data is discussed and put forward, and the attribute will be regarded as unqualified while the rate of disfigurement exceeds the quality tolerance. Besides,a study case is given to explain the practical application with the quality control method based on rate of disfigurement, ensuring the attribute data quality control effectively.更多还原