Object classification prediction method and device, equipment, medium and program product

The invention discloses an object classification prediction method and device, equipment, a medium and a program product, and relates to the field of artificial intelligence. The method comprises the steps of obtaining multiple pieces of historical behavior data and model parameters of multiple candidate models; determining centroid feature representations corresponding to the at least two classification categories based on the historical behavior data; performing parameter fusion on the model parameters of the plurality of candidate models based on the centroid feature representation corresponding to the ith classification category to obtain a first parameter corresponding to the ith classification category; converting the feature distribution condition represented by the centroid feature representation corresponding to the ith classification category into a second parameter corresponding to the ith classification category; and performing parameter fusion on the first parameter and the second parameter to ob