Article recall method and device, storage medium and electronic equipment

Abstract

The invention belongs to the technical field of computers, and relates to an article recall method and device, a storage medium and electronic equipment. The method comprises the steps of obtaining user data features, article information features and article category identifiers, and performing sparse mapping processing to obtain a to-be-input user vector; performing iterative full connection processing on the to-be-input user vector to obtain a first feature vector, and performing feature normalization processing to obtain a user representation vector; performing sparse mapping processing on the article information features to obtain a to-be-input article vector, and performing iterative full connection processing to obtain a second feature vector; performing feature normalization processing on the second feature vector to obtain an article representation vector, and performing category alignment loss calculation to obtain a category alignment loss value; and based on the category alignment loss value, performing dot product similarity calculation on the user representation vector and the article representation vector to obtain an article recall score so as to determine a recalled article. According to the invention, the recall accuracy of small-category articles and the recall precision of all categories of articles are improved.

Description

technical field [0001] The present disclosure relates to the field of computer technology, and in particular to an item recall method, an item recall device, a computer-readable storage medium, and electronic equipment. Background technique [0002] Recommender systems are widely used in e-commerce scenarios, which can filter items that users are interested in from a large number of candidate items. The core challenge of this large-scale recommendation is to accurately score tens of millions or even hundreds of millions of items in real time. A common practice in the industry is to view recommender systems as two-stage architectures. Among them, the recall module first recalls a small number of relevant candidate items from a large item corpus, and the sorting model sorts the recalled items according to the previously collected user historical behavior data. Obviously, the quality of the item candidate set recalled by the recall module will affect the effect of the subsequ...

AI Technical Summary

Problems solved by technology

In this way, the recall model will be inclined to large categories of items, which will make the recall accuracy rate of the largest item category the highest, and the recall accuracy rate of other categories of items will be low, which will indirectly damage the recall accuracy rate of other item categories.

Images