Techniques for identifying potential solutions for abnormal events based on historical data are described. One example method includes obtaining interaction data between a client device and a server; determining whether an abnormal event exists in the obtained interaction data; in response to determining, obtaining attribute information of a user corresponding to the client device from the obtained interaction data; identifying one or more records in a database, each record including the attribute information, a cause of one or more previous abnormal events, and a weight characterizing a severity of the previous abnormal events; determining a cause of the abnormal event, wherein the determined cause is the same as the cause of the previous abnormal events from the record in the identified one or more records that includes a largest weight; and providing a solution to address the abnormal event based on the determined cause.