Continuous audit early warning method and system

Abstract

The invention relates to a continuous auditing early warning method, which is applied to an auditing science and technology platform, and comprises the following steps: S1, carrying out self-service modeling through an auditing graphical self-service modeling engine, generating an SQL script, synchronously transmitting the SQL script to a data lake for execution, returning a result, and establishing an auditing early warning model; s2, performing continuous audit early warning parameter threshold configuration according to the indexes determined by the audit early warning model; s3, synchronously transmitting the audit early warning model and the continuous audit early warning parameter threshold configuration information to a data lake for calculation to obtain first continuous audit early warning data; s4, the data lake performs offline processing on the first continuous auditing early warning data to obtain second continuous auditing early warning data; and S5, the data lake synchronously transmits the second continuous auditing early warning data to the auditing science and technology platform, and early warning prompting and pushing are completed on the auditing science and technology platform. And the effects of supporting more than one billion level of a single table, totally storing more than 2PB of flexible and continuous audit early warning data operation and fast drilling check at millisecond level are realized.

Description

technical field [0001] The invention relates to the field of big data technology, in particular to a continuous audit early warning method and system. Background technique [0002] In the existing audit system of a certain bank, the continuous audit early warning is to process the detailed data early warning based on the enterprise data warehouse or audit system database through the development of the rule model, and provide the detailed query function. [0003] In the actual operation process, there are still the following problems: [0004] 1. The traditional mpp data storage cannot meet the storage and computing power needs of all the data of the entire bank, and the hardware cost is high, the linear expansion capability is poor, and the computing power cannot meet the business development of the whole bank. The query response efficiency of tens of billions of data is low, and it cannot Realize full-text fuzzy query, and cannot support the storage of detailed business da...

AI Technical Summary

Problems solved by technology

[0004] 1. The traditional mpp data storage cannot meet the storage and computing power needs of all the data of the entire bank, and the hardware cost is high, the linear expansion capability is poor, and the computing power cannot meet the business development of the whole bank. The query response efficiency of tens of billions of data is low, and it cannot Realize full-text fuzzy query, unable to support full-line detailed business data storage

[0005] 2. Real-time query of the head office-first-tier branch-second-tier branch-first-tier branch-outlet-transaction details and suspicious data level-by-layer drill-down cannot be achieved at a data level of more than 10 billion, and there is a lack of macro-level indicator drill-down at the institutional level , it is impossible to drill down to the detailed level of massive business data through various indicators of the organization

[0006] 3. The update of the continuous audit early warning rules cannot be synchronized to the data warehouse in time, the timeliness is poor, and the early warning adjustment rules are not flexible

[0007] 4. Due to the characteristics of the data structure and technical structure, the continuous audit early warning cannot be pushed to the front, stage, and middle stage departments of the enterprise to deal with the early warning. There is a lack of three lines of defense joint defense and joint control mechanism, and the early warning information cannot be quickly and effectively implemented and rectified

Images