Method for recommending students for topic under perspective of heterogeneous information network based on graph embedding

• Main Authors: WANG YIFENG, WU ZEZHENG, ZHANG JINGWEI, JIANG HAONAN, WANG DUIQIANG
• Format: Patent
• Language: Chinese; English
• Published: 02.12.2022
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FORADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORYOR FORECASTING PURPOSES; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE,COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTINGPURPOSES, NOT OTHERWISE PROVIDED FOR

The invention discloses a method for recommending students for a subject under the perspective of a heterogeneous information network based on graph embedding. The method comprises the following steps: preprocessing data; constructing a heterogeneous information network graph; acquiring information sequences of different nodes through a random walk algorithm under the heterogeneous information network diagram; carrying out position information modeling on nodes in the sequence, helping a model to understand the meaning of the current node in the sequence through a multi-head self-attention layer, carrying out residual connection and feedforward network operation on a result of the multi-head self-attention layer, and outputting a vector weight after a subject node absorbs information in the sequence; constructing positive and negative samples between the subject and the students, and performing feedback training on the model through loss calculation to obtain an optimal model; according to the method, the com