Movie personalized sequence recommendation method based on deep clustering algorithm

The invention discloses a movie personalized sequence recommendation method based on a deep clustering algorithm, which realizes personalized movie recommendation. The method comprises the following steps: acquiring a movie data set, and separating required data; mining a high-level intention of the user from the obtained data through deep clustering to form an interaction sequence fusing the high-level intention of the user; performing user static feature mining on the obtained data, and combining with a user film watching sequence to form a user feature dynamic sequence; combining the interaction sequence fused with the high-level intention of the user and the dynamic sequence of the user characteristics, and inputting the combination into a gating circulation unit to obtain fused sequence information; and calculating matching scores of all candidate items and the sequence by using the fused sequence information, and selecting the top K items from the candidate items as recommendation results. According to