Model search space generation method, device and system

Abstract

The embodiment of the invention provides a model search space generation method, device and system. The method comprises the steps of obtaining modeling task information, wherein the modeling task information comprises target data set information; processing the feature information of the data set by using a pre-trained recommendation model for a recommendation model training method to obtain a recommendation result; generating a model search space according to the recommendation result; wherein the recommendation result comprises a feature engineering strategy and a model training algorithm, and the model search space comprises an operator module space and an operator parameter space. According to the embodiment of the invention, the model search space generated according to the recommendation result has convergence, and massive calculation does not need to be performed on the model search space, so that the resource consumption of a computer for model training is reduced. In addition, due to the fact that the model search space has convergence, the time for the modeling tool to select the feature processing strategy, the model algorithm and the corresponding hyper-parameters from the model search space is greatly shortened, and therefore the parameter search efficiency in the model training process is improved.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, in particular to a method, device and system for generating a model search space. Background technique [0002] With the improvement of social informatization and intelligence, using modeling tools to train business models and using trained business models to realize intelligent processing of big data services has gradually become a common method in the big data industry. [0003] In the training process of existing models, a huge search space is usually defined, which usually includes a variety of processing feature strategies, model algorithms and corresponding hyperparameters. The modeling tool searches the feature processing strategy, model algorithm, and corresponding hyperparameters from the search space for model training. However, because the search space is very large, the process of selecting feature processing strategies, model algorithms, and corresponding hyp...

AI Technical Summary

Problems solved by technology

[0004] Embodiments of the present invention provide a method, device, and system for generating a model search space, which can solve the time-consuming problem that existing modeling tools select feature processing strategies, model algorithms, and corresponding hyperparameters from the search space. The computer used for model training needs to consume a lot of resources

Images