In order to deal with the lack of information sharing and resource re-utilizations,a novel personalized recommendation model was presented based on multi-scene fusion,and its function units,working flows and the scene data structure were given.The model utilized a distributed bidirectional description method to process scene data,and multi-scene fusion algorithms were used to exchange recommendation information between client characters and service scenes.Simulation results showed that the model has better coverage rates,recommendation precisions and resource consumption rate than the absolute node algorithms.
