Abstract: To solve the problem of poor classification adaptability in traditional data credibility evaluation models, a data credibility evaluation model based on exponential hierarchical structure algorithm is designed. This paper analyzes the actual data asset management process, establishes the data credibility evaluation index system; supplies the missing data in the data to be evaluated according to the data type and the periodical relationship between the data, and completes the pretreatment of the estimated data; generates the data set after data normalization, establishes the sub-hypermetric space according to the correlation coefficient between the data, and generates the exponential hierarchical structure tree. The reliability model is designed by combining the analytic hierarchy process. The experimental results show that compared with the traditional evaluation model, the proposed model has better classification adaptability, higher data recall rate and more obvious application advantages.

Key words: exponential hierarchical structure algorithm, evaluation of data reliability, hierarchical analysis, missing data, sub-hypermetric space