Simulation Design of Product Portfolio Optimization Decision-Making Marketing Model Based on Collaborative Filtering Recommendation Algorithm

With the increasing competition in the market, retail enterprises need to find new profit growth points in order to gain competitive advantage. Through collaborative filtering recommendation algorithm, valuable knowledge can be extracted from a large number of accumulated data to guide the marketing strategy of enterprises. Product mix optimization decision-making is an important aspect of enterprise marketing decision-making. This kind of decision-making should not only consider the cost input and profit income of the enterprise, but also consider the constraints of various resource elements of the enterprise. In the inventory pledge financing business, in order to effectively control the risks, it is necessary to set up a reasonable portfolio pledge rate. This paper comprehensively measures the comprehensive risk control target and the pledge rate target under the conditions of different demand and pledge rate of portfolio pledge, and compares and analyzes the performance and applicability of portfolio pledge loan and single pledge loan. In order to solve the problem of data sparsity, the local optimization method is used to select the recommendation group when selecting the nearest neighbor object of the target, which reduces the prediction error, ensures that the error value converges to a fixed value, and adopts the content-based method to adjust the abnormal prediction results of collaborative filtering prediction. Experimental results show that the proposed algorithm suppresses the adverse effects of data sparsity and improves the accuracy of collaborative filtering prediction score.