TRAINING METHOD, DEVICE, EQUIPMENT AND STORAGE MEDIUM OF ONLINE PREDICTION MODEL

摘要

The present invention discloses a training method, an apparatus, a device, and a storage medium of an online predictive model, and relates to the field of machine learning technology. One specific embodiment of the method includes obtaining an offline sample feature and an online sample feature of a user, the offline sample feature comprising a user portrait feature; Performing offline training based on the user's offline sample characteristics and online sample characteristics to obtain an offline recommendation model; Acquiring the latest online features of the user, and performing online training based on the latest online features of the user to obtain an online learning model-The online learning model has the latest online features as an online sample feature, and has already been trained offline recommendation. It is used to adapt to the input to the model -; And synchronizing the offline recommendation model online, inputting the latest online features output by the online learning model into the offline recommendation model, and generating an online prediction model, whereby online prediction using the user's online sample data By optimizing the model, the recommended model fitting effect and the generalization effect are improved, and it can be applied to a cloud computing scenario, but is not limited thereto.