Commodity recommendation method based on neighbor users and comment information

The invention discloses a commodity recommendation method based on neighbor users and comment information, and the method comprises the steps: carrying out the data cleaning of the score evaluation data of a user for a commodity in a shopping platform, obtaining the pre-processed score classification and comment data of the commodity, and through a PV-DM model in a Doc2vec method, according to the cosine similarity, obtaining neighbor users and neighbor commodity auxiliary modeling user preferences and commodity attributes, constructing the training data, inputting the training data into a recommendation model based on a bidirectional long and short memory network, an attention mechanism and high-order feature intersection for training and commodity score prediction, and finally performing the commodity recommendation according to the predicted score. According to the method, user preferences and commodity attributes are modeled more accurately by introducing neighbor users and neighbor commodities based on c