A method and device for identifying users with abnormal behavior

Abstract

The invention discloses a method for identifying users with abnormal behaviors, which includes: obtaining training signaling data and testing signaling data from signaling data of a full amount of bills according to a black number database and a white number database; obtaining characteristic data in the training signaling data The first signaling data whose value is highly correlated with the eigenvalue of the target; generate a first information table including derived indicators according to the eigenvalue and time granularity of the first signaling data; use the first screening strategy to filter the derived indicators, generating a second information table including significant indicators; combining the first information table and the second information table to generate a training information table; establishing a first model based on the training information table, and using the test signaling data to A model is tested to obtain test results; the first model is evaluated according to the test results to obtain an evaluated and optimized second model to identify users with abnormal behaviors. The invention also discloses a device for identifying users with abnormal behavior. Using the method and device for identifying users with abnormal behavior provided by the present invention can reduce the adjustment and optimization time after the algorithm model goes online.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a method and device for identifying users with abnormal behavior. Background technique [0002] In the prior art, when identifying abnormal calling users or fraudulent numbers, generally first establish or collect blacklists through user self-marking, blacklists provided by the police and the Ministry of Information Security, third-party friendly business channels, etc., classify the blacklists, establish a database, and then The current number is compared with the database to identify the current number. This method cannot effectively detect users with abnormal behaviors or numbers involved in fraud in real time. [0003] In another scheme for identifying abnormal calling users or fraudulent numbers, the algorithm is used to analyze user behavior and form an algorithm model. This scheme can test whether the algorithm model is correct through existing data, and gradually after...

AI Technical Summary

Problems solved by technology

This method cannot effectively detect users with abnormal behaviors or numbers involved in fraud in real time

[0003] In another scheme for identifying abnormal calling users or fraudulent numbers, the algorithm is used to analyze user behavior and form an algorithm model. This scheme can test whether the algorithm model is correct through the existing data, and gradually after the algorithm model is actually launched. Adjustment and optimization, but it takes a long time for the algorithm model to go online to run stably, which makes it impossible to identify users with abnormal behavior for a long time

Images