Adaptive adjusting method and system of machine learning algorithm

Abstract

The invention relates to the adaptive adjusting method and system of a machine learning algorithm. The method comprises the following steps of through analyzing the machine learning algorithm, acquiring a controllable parameter used for controlling calculation time, and according to the actual calculation time of the machine learning algorithm under each specific value of the controllable parameter, establishing the quantitative model library of the calculation time of the machine learning algorithm; and according to the complexity of input data in each time window, carrying out coarse grain adjusting on the machine learning algorithm, giving the complexity range of an algorithm model, carrying out quantitative description on the input data according to the machine learning algorithm, determining the specific value of the controllable parameter in the quantitative model library through combining a given time limit, and applying the specific value to the machine learning algorithm so asto realize the adaptive adjusting of the machine learning algorithm. In the invention, the machine learning algorithm can adapt to the on-line application scene of streaming data under a limited calculating resource condition, and a reasonable calculating result can be given under a given time limit.

Description

technical field [0001] The invention relates to the fields of machine learning algorithm analysis and algorithm adaptive adjustment, and relates to an information entropy-based adaptive adjustment method and system for machine learning algorithms. Background technique [0002] Machine learning algorithms have been widely used in data analysis and data mining, and have achieved good results. At the same time, the machine learning algorithm is also a highly specific algorithm for specific applications or specific data, that is, once the training is completed and uniquely determined using the given training data, it will be difficult for the model to have a generalized performance for different data or different applications. ability. Therefore, in order to make the model adaptable to different data or applications, for example, by adjusting some controllable parameters in the model to automatically adjust its model structure, it is a very effective solution to expand the gene...

AI Technical Summary

Problems solved by technology

[0005] In the scenario of online application of streaming data, traditional machine learning algorithms continue to increase the complexity of these model structures in order to pursue the accuracy of model algorithms, thereby introducing additional computing time overhead—this brings real-time demand for online great shock

Images