Search data mining method and device, equipment and storage medium

Abstract

The invention relates to the technical field of natural language processing, in particular to a search data mining method and device, equipment and a storage medium. The method comprises the following steps: acquiring historical search behavior data; constructing a relation network diagram based on the historical search behavior data; obtaining a plurality of node sample pairs based on the relation network diagram; training a graph neural network model by using the plurality of node sample pairs to obtain word embedding vectors of all nodes in the relation network graph, wherein the graph neural network model comprises a graph neural network layer, and the graph neural network layer is used for capturing an adjacent node set corresponding to nodes in each node sample pair; removing other nodes in the node sample pair from the adjacent node set to obtain a target node set; performing feature aggregation on the nodes based on the target node set; and according to the word embedding vector, matching each node in the relation network diagram with other nodes to obtain a matching result.

Description

technical field [0001] The present invention relates to the technical field of natural language processing, in particular to a search data mining method, device, equipment and storage medium. Background technique [0002] A search engine refers to a system that collects information on the Internet by using a specific computer program according to a certain strategy, organizes and processes the information, and displays the processed information to users. It is a system that provides users with retrieval services. Vertical search (also known as special search) is the subdivision and extension of search engines, mainly for the search of a certain field, which is to integrate a certain type of special information in the information database, and extract the required data by directional sub-fields. After processing, it returns to the user in some form. Vertical search includes music search, image search, video search, official account search, and mini program search. [0003] ...

AI Technical Summary

Problems solved by technology

However, the accuracy and generalization ability of the model in the prior art is poor, which leads to the poor correlation between mining the relationship between query and doc and the relationship between query and query, which affects the effect of query recommendation and user experience. poor

Images