Variable discretization control method and system

Abstract

The invention discloses a variable discretization control method and system. The control method comprises the steps that: S1, variables are acquired; S2, discretization processing is conducted on thevariables; and S3, after the binning tangent points of the discretized binning of the variables are obtained, binning and encoding are carried out according to the tangent point variables. The step S2specifically comprises: obtaining an initial variable tangent point to form N candidate tangent points; defining genetic algorithm individuals according to a certain probability; on the basis of theinitial population, generating a next-generation population through evolutionary operation; in the process, gradually eliminating individuals with numerical values lower than a set threshold value, and increasing individuals with high fitness function values; after a plurality of generations are evolved, enabling the individual with the highest fitness function value to be the selected optimal individual vector, and enabling the segmentation point of the optimal individual vector to be the optimal tangent point of the variable; and taking the segmentation point of the optimal individual vectoras a binning tangent point of discretization binning. According to the variable discretization control method and system, the evolution time can be shortened, and the data processing speed can be increased.

Description

technical field [0001] The invention belongs to the technical field of variable processing, and relates to a variable processing method, in particular to a variable discretization control method and system. Background technique [0002] In order to solve the problem of over-fitting or under-fitting caused by large variation range of feature data and many outliers in machine learning modeling, it is necessary to discretize continuous feature data. Genetic Algorithm is an algorithm for finding the optimal solution based on the theory of biological evolution. It simulates the problem to be solved as a process of biological evolution, and follows the principle of "survival of the fittest" to evolve the optimal individual. Discretization can be regarded as the problem of selecting the optimal segmentation point, which can be transformed into an optimization problem by combining the genetic algorithm to select the appropriate fitness function. [0003] In the existing practical a...

AI Technical Summary

Problems solved by technology

The application of the existing genetic algorithm in discretization is generally given the number of tangent points, and the IV value is used as a fitness function. It works well for variables with a high degree of dispersion, but it is not effective for continuous values ​​​​with very fine granularity.

[0004] The genetic algorithm has only one framework in the application of discretization, and there is no unified specific plan for the definition of the individual, the design of the fitness function, or the initialization method of the initial population.

Especially in terms of how to consider both the discretization effect and the time complexity, the existing technology has not given an objective conclusion

Images