MESSAGE-PUSHING METHOD AND APPARATUS BASED ON IMPLICIT MULTI-TARGET FUSION OF MODELS

Disclosed in the present invention are a message-pushing method and apparatus based on implicit multi-target fusion of models. The method comprises: acquiring sample data of each target sub-process for message pushing, wherein the target sub-processes are arranged in a chronological order, and each target sub-process is a posterior process of a previous target sub-process and/or an anterior process of a following target sub-process; training a first model according to sample data of the posterior process, training a second model according to sample data of the anterior process, and implicitly fusing sample information of the first model into the second model; inputting prediction data of the anterior process into the second model to obtain a prediction result; and performing message pushing according to the prediction result. The present invention can avoid the problem of samples among a plurality of service targets being imbalanced, realizes end-to-end prediction, and reduces the complexity of a pushing syst