Object recommendation method and device, model training method and device, equipment, medium and product

The invention relates to a service object recommendation method and device, and a model training method and device, and the method comprises the steps: obtaining training sample data, wherein the service object value of positive service object sample data is greater than the service object value of negative service object sample data; and according to the user information sample data, the positive service object sample data and the negative service object sample data, respectively performing mapping to obtain the probability of the positive service object sample data selected by the user and the probability of the negative service object sample data selected by the user, and performing training in combination with constraint conditions to obtain a service object recommendation model. According to the method and device, the influence of absolutely large or absolutely small inaccuracy of the service object value on model training is weakened by utilizing the relative value of the service object value, the proba