Entity transaction attribute determination method and apparatus

An entity transaction attribute determination apparatus, comprising a processor and a non-transitory computer-readable memory coupled to the processor, and configured with instructions executable by the processor to perform operations, comprising:

• obtaining a plurality of historical relational networks sequentially arranged under a temporal order, wherein the historical relational networks respectively correspond to historical time points having a preset time interval, each of the historical relational networks comprises a plurality of nodes, each node corresponds to a plurality of entities comprising a to-be-predicted entity, each of the historical relational networks is configured to describe association relationships between the plurality of entities corresponding to a corresponding historical time point, and each node corresponds to a node vector composed of transaction characteristics extracted based on descriptive information of the corresponding entities;
• determining, for each of the historical relational networks and through vector fusion of neighbor nodes, a plurality of description vectors of the to-be-predicted entity;
• processing, through a pre-trained time-series neural network, the description vectors to obtain an output result;
• and
• determining, according to the output result, an attribute state or an attribute category of the preset transaction attribute for the to-be-predicted entity.

A non-transitory computer-readable storage medium having instructions stored thereon executable by a processor to cause the processor to perform operations comprising:

• obtaining a plurality of historical relational networks sequentially arranged under a temporal order, wherein the historical relational networks respectively correspond to historical time points having a preset time interval, each of the historical relational networks comprises a plurality of nodes, each node corresponds to a plurality of entities comprising a to-be-predicted entity, each of the historical relational networks is configured to describe association relationships between the plurality of entities corresponding to a corresponding historical time point, and each node corresponds to a node vector composed of transaction characteristics extracted based on descriptive information of the corresponding entities;
• determining, for each of the historical relational networks and through vector fusion of neighbor nodes, a plurality of description vectors of the to-be-predicted entity;
• processing, through a pre-trained time-series neural network, the description vectors to obtain an output result;
• and
• determining, according to the output result, an attribute state or an attribute category of the preset transaction attribute for the to-be-predicted entity.