Research And Implementation Of Data Quality Evaluation Algorithm Based On Metadata

The level of data quality plays a key role in enterprises and even countries.The possession of high-quality data has become an important guarantee for the development of enterprises.How to check the quality of data and how to find out the problems in data have become a hot issue in the information age.Data quality assessment is the effective way to measure data problems.we can know the quality of data and the location of problem data by evaluating the quality of data,which lays the foundation for improving the quality.The dimensions of data quality assessment include integrity,consistency,accuracy,relevance,timeliness and so on.Usually,the assessment of data quality is carried out through multiple dimensions and the determination of these dimensions needs to be determined by the characteristics of data itself.For the quality of data in relational databases,this paper puts forward a universal rules extraction model based on metadata,which solves the problem of using a single dataset for quantitative evaluation.The model can carry on the unified handling and extract metadata to the data source and some assessment rules.By establishing interface for heterogeneous data source,can we obtain the metadata rule base from analyzing the database,data tables to data items deeply.The rule base provides the basis for data quality assessment.Evaluation algorithms are set up on the integrity,accuracy and consistency dimensions.Finally,the entire database is evaluated.Evaluation results can reflect the status of the quality of the data clearly,showing that the algorithms under the model are practicability and reliability.