Data prediction system and method based on two-layer model structure

Abstract

The present invention provides a data prediction system and method based on a two-layer model structure, including a model configuration management unit, a model execution management unit, and a model container management unit; the model configuration management unit includes a dedicated model configuration management subunit and a general model configuration management subunit Unit, general model library, special model library and model information service subunit; model container management unit includes model container scheduling subunit, dedicated container resource pool, basic container resource pool and model container monitoring subunit. The advantages are: (1) Through the two-layer model structure, the reusability of general models can be greatly improved, thereby shortening the development cycle of complex prediction models in various business systems. (2) By adopting the isolation operation technology method of the container, the normal, stable, safe and reliable operation of the two-layer model structure is guaranteed, and the conflict of the execution of the special model of different users is avoided, resulting in abnormal execution.

Description

technical field [0001] The invention belongs to the technical field of data prediction, and in particular relates to a data prediction system and method based on a two-layer model structure. Background technique [0002] Data prediction technology is widely used in various decision support systems. In the prior art, the method of data prediction is generally as follows: According to a specific application requirement, for example, for the price prediction demand of meat products, or for the demand of vegetable price prediction, respectively write corresponding data prediction display units, including: writing data Predict the model, write the data analysis code, write the result display code, and then realize the data prediction by executing the data prediction display unit, and display the data prediction result. [0003] The above data prediction method has the following problems: Since each specific application needs to write a set of data prediction display unit, the re...

AI Technical Summary

Problems solved by technology

[0003] The above data prediction method has the following problems: Since each specific application needs to write a set of data prediction display unit, the reusability of each data prediction display unit is very low, which increases the workload of programmers and has the advantages of data prediction. The development cycle of the display unit is long; in addition, the scalability of the data prediction display unit is not limited. Whenever the function needs to be expanded, it needs to be improved from the program code level, which reduces the efficiency of the data prediction display unit expansion.

Images