BDMCA: a big data management system for Chinese auditing.

Area of Science:

• Computer Science
• Data Management
• Auditing

Background:

• Big data technologies offer opportunities for Chinese audits but face challenges like data heterogeneity and inefficient processing.
• Existing methods struggle with multi-source data integration and converting Chinese queries into SQL.
• These obstacles hinder the effective application of big data in Chinese auditing.

Purpose of the Study:

• To propose BDMCA, a big data management system tailored for Chinese audits.
• To develop solutions for handling heterogeneous audit data and improving data processing efficiency.
• To enhance the accuracy and performance of Chinese audit data analysis.

Main Methods:

• Developed a hybrid management architecture to address multi-mode data heterogeneity.
• Defined an R-HBase spatio-temporal meta-structure for efficient auditing.
• Utilized slot value filling for template generation and created the MRo-SQL learning model.

Main Results:

• R-HBase demonstrated significant improvements over MD-HBase in range and kNN queries (4.5x and 3x faster, respectively).
• MRo-SQL achieved superior logical-form accuracy (up to 5.2%) and execution accuracy (up to 5.9%) compared to X-SQL.
• The hybrid architecture effectively alleviated data heterogeneity issues.

Conclusions:

• BDMCA provides an effective solution for managing big data in Chinese audits.
• The proposed R-HBase structure and MRo-SQL model significantly enhance data processing and query capabilities.
• These advancements pave the way for more efficient and accurate Chinese audit practices leveraging big data.