Reliability Analysis and Data Snooping

Reliability analysis and data snooping are important contents in quality control of adjustment system.Baarda's reliability analysis and data snooping method is designed for independent observations and it is quite consummate both in theory and in practice.But they are researched respectively and constitute different systems.In this paper,we establish a uniform basic equation used for reliability analysis and data snooping from verifying the ratio of variance,then the expression of reliability index based on the correlation weight is obttained.We also prove that the expression applied to find minimal model errors and the weight matrix is not restricted to diagonal matrix.When the weight matrix is diagonally,the redundant observations' inner reliability index r_i is a particular case of R_i.The consistency proves that R_i is reasonable.At the same time,a new data snooping method is presented,which is related to reliability analysis measurement.The consistency of reliability analysis and data snooping both in theory and in practice is a new viewpoint,which will be used in other fields of surveying data processing.