Data-driven complex system mechanism automatic learning method, system and equipment

Abstract

The invention belongs to the field of big data and machine learning, particularly relates to a data-driven complex system mechanism automatic learning method, a system and equipment, and aims to solve the problems that an existing system modeling technology is difficult to predict a behavior trend from field observation data, a reconstructed mechanism model is not matched with physical observation data, the robustness is poor and the like. The method comprises the steps of obtaining historical multi-modal data and real-time multi-modal data, constructing a time sequence long-range correlation hypergraph model through airspace circulation memory coding, performing normalized combination on the hypergraph model through a neural differential equation model, and performing automatic iterative search on a continuous game network structure to obtain a system mechanism continuous dynamic model, and performing biological evolution simulation to obtain a causal model, and recalculating an association weight to obtain an active early warning system. According to the method, non-linearity, emergence, balance step, adaptability and special property description of a feedback loop of a complex system are realized, and the prediction accuracy of the model is improved.

Description

technical field

[0001] The invention belongs to the field of big data and machine learning, and in particular relates to a data-driven automatic learning method, system and equipment for complex system mechanisms. Background technique

[0002] With the rapid development of information technology, human production activities increasingly rely on various complex systems, such as smart manufacturing, the Internet, smart cities, smart medical care, energy Internet, smart transportation and ecosystems, etc. These complex systems are a whole formed by several units or subsystems through interrelationships. According to human wisdom, these systems are endowed with symbiosis and symbiosis. These complex systems continue to accumulate a large amount of dynamic multi-modal data, which contains valuable knowledge such as the working mechanism behavior, state and regulation optimization of the system from the micro to the macro. Machine learning plays an important role in the mechanism...

Images