Negative sample extraction method based on path semantics and feature extraction

A technology of feature extraction and extraction method, applied in the field of negative sample extraction based on path semantics and feature extraction, can solve the problems of small amount of information, useless features cannot be properly suppressed, etc., and achieve the effect of improving the representation ability

CN113342995AActive Publication Date: 2021-09-03CHENGDU UNIV OF INFORMATION TECH
7 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Current Assignee / Owner
Publication Date
2021-09-03

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention relates to the technical field of recommendation systems, in particular to a negative sample extraction method based on path semantics and feature extraction. The method comprises the following steps: S1, carrying out sample collection, collecting a sample range needing to be extracted, and building a sampling system model; S2, combining representation learning of a triple structure of the knowledge graph with SDAE, and then obtaining codes of article entities from the relation; S3, according to the incidence relation between user nodes and article nodes in the knowledge graph, obtaining a negative sample through sampling in combination with a corresponding search algorithm; S4, sorting the negative sample data sets according to scores; S5, trimming the low-score negative sample data sets; and S6, carrying out a contrast test based on the data set. According to the method, an existing model is optimized by adopting a path-based comprehensive mode, starting from a positive sample, node relations in a spectrum are explored in a recursive manner through a reinforcement learning method, useful features are enhanced in combination with channel attention, and small-information-amount or useless features are properly inhibited.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention relates to the technical field of recommendation systems, in particular to a negative sample extraction method based on path semantics and feature extraction. Background technique

[0002] With the rapid development of computer technology, the wide use of various industries and APPs has generated a large amount of data, but not all of the data are of interest to users. Therefore, the recommendation system has emerged as the times require and has become an important technology to solve this problem. How to improve user satisfaction and experience and ensure the accuracy of recommendations has become the main research content of recommender systems. According to the horizontal speculation of user hobbies and the longitudinal analysis of user historical choices, recommendation systems can be divided into two categories: collaborative filtering-based recommendation systems and content-based recommendation systems. The former mainly faces the...

Examples

Embodiment Construction

[0017] In order to make the purpose, technical solution and advantages of the present invention clearer, the present invention will be further elaborated below in conjunction with the accompanying drawings. In describing the present invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", " The orientation or positional relationship indicated by "outside", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the referred device or element must have a specific orientation, so as to Specific orientation configurations and operations, therefore, are not to be construed as limitations on the invention.

[0018] The present invention provides a negative sample extraction method system based on path semantics and feature extraction. The system i...