Optimization method based on random walk relationship discovery

Abstract

The invention discloses an optimization method based on random walk relationship discovery, belongs to the technical field of knowledge maps, and relates to a relationship discovery problem in the knowledge map. The invention provides the random walking optimization method under a defined radius based on the random walk relationship discovery. The method is characterized in that aiming at a largescale knowledge base, the magnitude of knowledge map nodes is excessive along with the base, a local submap obtaining the defined radius under a breadth-first search is combined, and a large amount oftime and storage space consumed when preprocessing is conducted are accelerated by adopting the obtained local submap but not the whole map. Compared with a tradition random walk algorithm, random walking is conducted under the local submap so that the preprocessing time and the magnitude of space complicity can be greatly reduced to shorten random walking time on the basis of ensuring the accuracy of the relationship discovery.

Description

technical field [0001] The invention relates to an optimization method based on random walk relationship discovery, which belongs to the technical field of Knowledge Graph. Background technique [0002] Relational discovery in knowledge graphs, also known as relational reasoning, refers to starting from the existing relationships in the knowledge base and finding potential relationships through computer algorithms and programs, thereby further expanding the knowledge graph. Relationship discovery is very important in the process of creating and expanding knowledge graphs. At present, there are mainly two methods for relationship discovery in knowledge graphs: logic-based reasoning methods and graph-based reasoning methods. [0003] Logic-based relational reasoning mainly includes first-order predicate logic, description logic and rule-based reasoning. The first-order predicate logic is based on propositions. In the concept of first-order predicate logic, propositions are ...

AI Technical Summary

Problems solved by technology

[0008] The purpose of the present invention is to solve the problem that the existing random walk relationship discovery algorithm of knowledge graph has too many levels of time and space complexity when dealing with large-scale knowledge bases. The existing random walk relationship discovery algorithm of knowledge graph In the process of calculating the iterative convergence results of the affinity of the random walk, the algorithm needs to calculate the inverse matrix of the square matrix of the node order of the knowledge graph. As the scale of the knowledge base continues to increase, the overhead of the calculation will also increase. Rapidly increasing, this invention is dedicated to reducing the time and space overhead in the process of large-scale knowledge map relationship discovery, speeding up the process of relationship discovery, and proposes an optimization method based on random walk relationship discovery

Images