An optimal matching algorithm for e-commerce recommendation information based on matrix decomposition

Aiming at the problems of data sparsity and cold start in traditional e-commerce recommendation information matching, an optimal matching algorithm for e-commerce recommendation information based on matrix decomposition is proposed. On the basis of e-commerce behaviour, the optimal user behaviour is obtained by decomposing the user's behaviour and the user's preference, and then the optimal user behaviour is obtained by decomposing the user's behaviour and the user's preference. The experimental results show that compared with the traditional optimal matching algorithm for recommendation information, the proposed optimal matching algorithm for e-commerce recommendation information based on matrix decomposition has a higher area under curve (AUC) value and a lower root-mean-square error (RMSE) value, and the performance of recommendation information matching is better.