Commodity sequence recommendation method based on attention mechanism

The invention discloses a commodity sequence recommendation method based on an attention mechanism, and the method comprises the steps: obtaining a commodity historical data set of a user, and carrying out the data preprocessing; a multi-head attention mechanism is used for capturing relation dependence between articles in recent historical behaviors of the user, and a short-term interest preference vector of the user is obtained; and adding the similarity coefficient of the user and the item in the implicit space into the weight of the attention mechanism, and distributing different weights for the long-term interest to obtain the final representation of the long-term preference. And finally, performing weighted fusion on long and short term preferences to obtain final interest preferences of the user. The method comprises the following steps of: acquiring a correlation score value, then interacting the correlation score value with each article in a candidate set, acquiring the correlation score value in a p