The invention provides a user behavior
risk control method and
system based on an OTA platform, and relates to the technical field of
Internet service security, and the method comprises the steps: collecting equipment information, interaction behavior sequence information and service operation information of a user in real time, and constructing a multi-
modal behavior
feature vector; and generating a behavior event vector based on the interaction behavior sequence information and the service operation information, determining a dynamic weight vector based on the multi-
modal behavior
feature vector, obtaining a historical reputation
score of the user, and updating the historical reputation
score based on the weight vector and the event vector to obtain a current dynamic reputation
score. And carrying out
risk assessment on the current dynamic reputation score, obtaining a
risk assessment result and executing a corresponding hierarchical
risk control disposal strategy, thereby solving the problems that an existing
risk control system is easy to be evaded by black products due to dependence on a static
single rule, the accidental injury rate of a public scene is high,
context awareness and self-evolution ability are lacked, and the contradiction is poor. The method has the effects of improving black product
behavior recognition accuracy and ensuring normal user experience.