Research On Comprehensive Evaluation Method For Database Digital Fingerprint

With the rapid development of information and communication technology,database is more and more widely used in various information fields.Database distribution is the most common application scenario,and ensuring data security during the process is of paramount importance.Therefore,database digital fingerprinting technology was born.In order to select a higher performance database digital fingerprint for a given database and a specific usage scenario,as well as to improve the theoretical system of database digital fingerprinting,it is of strong practical importance to conduct research on the comprehensive evaluation of database digital fingerprinting.In the current process of comprehensive evaluation of database fingerprints,the following problems mainly exist: lack of a unified and comprehensive index system and quantification methods,biased weight determination methods,and insufficient interpretability of the determined weights.To address the above problems,the main work of this paper is as follows.Firstly,this paper establishes a database fingerprint hierarchy evaluation index system for database use scenarios,combining the performance evaluation indexes of traditional digital fingerprints,and proposes a solution for quantifying the indexes and evaluating data pre-processing.The index system,while comprehensively and objectively describing the comprehensive performance of database digital fingerprints,clearly and accurately reflects the logical relationship between the indicators and lays the foundation for the comprehensive evaluation of database digital fingerprints.Then,to address the problems of subjectivity and calculation efficiency in the determination of database fingerprint index weights,this paper proposes a linear combination scheme of comprehensive subjective and objective weights,which reduces the influence of expert subjectivity and evaluation data itself in determining weights and improves the representativeness of weights.The extension superiority is used to calculate the correlation degree value of the object to be evaluated with different grades,which enriches the evaluation result data of the object to be evaluated and makes the performance comparable among different database digital fingerprints.Finally,in order to solve the problem of interpretability of combined weights,this paper proposes a comprehensive evaluation algorithm for database fingerprints based on ensemble classification method,using traditional evaluation algorithms to obtain evaluation results directly,and proposes a grading algorithm based on silhouette coefficient and K-mean clustering to solve the problem of inconsistent evaluation results between different evaluation algorithms and obtain grading evaluation results.The grading evaluation results and the ensemble classification algorithm are used to construct an ensemble classifier to realize the accurate grading and comprehensive evaluation of the digital fingerprints of the database.The experimental results show that: the proposed comprehensive evaluation algorithm based on the extension superiority degree has good usability,which can provide more accurate evaluation results,and provides richer evaluation results compared with similar algorithms;the proposed evaluation algorithm model based on ensemble classification has better accuracy and generalization ability than similar classifiers,and can effectively guide decision-makers to select database fingerprint with better performance.