Method and apparatus for optimizing generalized regression neural network by fruit fly algorithm

Abstract

Embodiments of the invention provide a method and an apparatus for optimizing a generalized regression neural network by a fruit fly algorithm. The method comprises the steps of obtaining sample data;according to the sample data, building a generalized regression neural network model; based on the generalized regression neural network model, determining an optimal propagation parameter value of the generalized regression neural network model by adopting the fruit fly algorithm, wherein the fruit fly algorithm has a progressively decreased fruit fly search step length value; and taking the optimal propagation parameter value as a propagation parameter value of the generalized regression neural network model, and obtaining an optimized generalized regression neural network model. The fruitfly search step length value of the fruit fly algorithm used for optimizing the generalized regression neural network model is progressively decreased, so that relatively speaking, the fruit fly search step length value during starting of search is relatively high; and therefore, the convergence in the beginning is quicker and the search efficiency is improved. The fruit fly search step length value during final search is relatively low, so that the search precision is improved.

Description

technical field [0001] The embodiments of the present application relate to the technical field of data processing, and in particular to a method and device for optimizing a generalized regression neural network with a fruit fly algorithm. Background technique [0002] Data mining is a step in database knowledge discovery, and generally refers to the process of searching information hidden in it through algorithms from a large amount of data. It can help decision makers extract potential relationships and patterns hidden in the data through the analysis of historical data and current data, and then assist them in predicting possible future conditions and upcoming results. Data mining has been widely used in the financial industry, retail industry, medical care, telecommunications and electric power and other fields, and has become an effective method and way to utilize information resources, and has broad development prospects and application markets. [0003] Taking short-...

AI Technical Summary

Problems solved by technology

When the fruit fly search step value is too large, the global optimization ability of the fruit fly algorithm is strong, but the local optimization ability is weak, so the accuracy is not high; when the fruit fly search step value is too small, the local optimization ability of the fruit fly algorithm The optimization ability is enhanced, but the global optimization ability is weak, and the algorithm is easy to fall into the local optimum

However, the fruit fly search step size in the fruit fly algorithm in the prior art is a fixed value, resulting in a weaker search ability of the fruit fly algorithm

Images