Model prediction result data management system and method

Abstract

The invention discloses a model prediction result data management system and method, and the system comprises: a result data storage module which is used for obtaining result data generated in a modelprediction process, and storing the result data in a preset path; a result data management module which is used for receiving a control instruction input in the result data set management page and processing the result data, wherein the control instruction comprises instructions including data storage, data preview, data maintenance, data deletion and data export, the result data is stored in a result data set according to a data storage instruction, and the result data is previewed, maintained or deleted according to a data previewing, data maintaining or data deleting instruction; a resultdata exporting module which is used for exporting corresponding result data according to the data exporting instruction and transmitting the result data to a data user in a file transmission mode; anda result data set which is used for storing result data generated in the model prediction process.

Description

technical field [0001] The invention relates to the technical field of model training, in particular to a system and method for managing model prediction result data. Background technique [0002] In the process of artificial intelligence modeling, after the model is often trained, the user needs to take the model away, and then perform the prediction task alone, so as to obtain the data he wants, and then make business adjustments based on the data. If after performing the prediction task, it is found that the effect of the model is not good, and the predicted result data is not the effect you want, then you need to retrain, and then take the model away again to perform the prediction task, and the development cycle and cost are high. [0003] To sum up, the existing technology takes away the model through modeling training, then executes the prediction task, and evaluates the result set data until the process meets the business requirements. With unified management, model...

AI Technical Summary

Problems solved by technology

If after performing the prediction task, it is found that the effect of the model is not good, and the predicted result data is not the effect you want, then you need to retrain, and then take the model away again for the prediction task, the development cycle and cost are high

[0003] To sum up, the existing technology takes away the model through modeling training, then executes the prediction task, and evaluates the result set data until the process meets the business requirements. Unified management, modelers cannot directly perform prediction tasks and save result datasets and evaluation result datasets, resulting in longer development cycles, higher development costs, and inability to control access rights

Images