Data exception analysis method and device and storage medium

Abstract

The invention provides a data exception analysis method and apparatus, and a storage medium. The method comprises the steps of obtaining historical operation data of a target operation behavior; performing reconciliation analysis processing on the historical operation data based on a reconciliation analysis model to obtain a reconciliation analysis result of the target operation behavior in targettime; wherein the reconciliation analysis model is obtained by performing machine learning training based on sample operation data; acquiring real operation data of the target operation behavior in the target time; and based on the real operation data and the reconciliation analysis result, determining an abnormal data analysis result of the target operation behavior in the target time. Accordingto the method, the data reconciliation efficiency and intelligence can be improved, the labor cost is reduced, the future data change trend is predicted according to the historical data, the abnormality in the data is actively found, and the data monitoring quality, efficiency and accuracy are improved.

Description

technical field [0001] The invention belongs to the technical field of data reconciliation, and in particular relates to a data anomaly analysis method, device and storage medium. Background technique [0002] Data reconciliation is mainly to check and verify the sending and receiving messages when data messages are circulated in the data system, so as to verify the correctness and reliability of data message circulation. [0003] The existing technology mainly adopts such as figure 1 As shown in the traditional reconciliation method based on rule strategy, the traditional reconciliation method first defines field rules, single-table rules, multi-table rules, etc. to form a rule template. Of course, it also supports Structured Query Language (SQL) or The script's self-defined rules, followed by the audit rules and reconciliation results, trigger corresponding alarms when the rules are met. [0004] Traditional reconciliation based on rule strategy, as long as the threshold...

AI Technical Summary

Problems solved by technology

However, the threshold needs to be specified in advance before the rules can be generated, and the threshold is often manually adjusted multiple times, which is less intelligent; and traditional reconciliation requires manual analysis of data to generate rules and achieve reconciliation, which has high labor costs and low reconciliation efficiency; In addition, traditional reconciliation cannot predict future data trends based on historical data. Actively discover anomalies in data (such as abnormal points and outliers)

Images