Cause inference regarding network trouble

Abstract

To accurately infer a cause of a trouble occurring in a network. A network system acquires first information indicating a status of the network (S), selects, from among a plurality of cause inference models provided for each type of a cause of a failure occurring in the network, each of the cause inference models being associated with a model selection condition, a cause inference model corresponding to the model selection condition satisfied by the acquired first information, the plurality of cause inference models being trained based on training data including input data that includes second information indicating the status of the network in a time period different from the time period of the first information and ground truth data indicating a cause of a trouble (S), acquires an output obtained when the first information indicating the status of the network is input to the selected cause inference model (S), and executes a process for responding to a trouble that has occurred in the network based on the acquired output of the selected cause inference model (S).