Method for predicating whether user has malicious behavior and calculation device

• Main Author: CAI ZIBIN
• Format: Patent
• Language: Chinese; English
• Published: 03.10.2017
• Subjects: CALCULATING; COMPUTING; COUNTING; DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FORADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORYOR FORECASTING PURPOSES; HANDLING RECORD CARRIERS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE,COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTINGPURPOSES, NOT OTHERWISE PROVIDED FOR

The invention discloses a method for predicating whether a user has a malicious behavior based on a user relation graph. The method is suitable for being executed in a calculation device. The user relation graph comprises nodes and edges for connecting associated nodes. The method comprises the following steps: obtaining user data of a user; updating the user relation graph according to the obtained user data; according to the user relation graph, judging whether the user is matched with preset abnormal modes; according to the matched abnormal mode and a probability value corresponding to the abnormal mode, calculating malicious probability of the user; and predicating that the user, the malicious probability of which is within a preset value interval, has the malicious behavior. The invention also discloses the corresponding calculation device and a computer readable storage medium. 本发明公开了种基于用户关系图来预测用户是否存在恶意行为的方法，适于在计算设备中执行，用户关系图包括节点和连接关联节点的边，该方法包括步骤：获取用户的用户数据；根据所获取的用户数据来更新用户关系图；根据用户关系图，判断用户是否与预置的异常模式相匹配；根据所