Research on a sample learning model based on SVR-GA hybrid algorithm

With the rapid development of industry, a lot of companies demand a high degree of material plate flatness. With the leveler in a certain machine structure, the selecting of leveler´s technical parameters directly decides the plate flatness after leveling. In this paper, through the analyses of samples´ features in the work of leveling and the study of related theories, combined with the requirements of its work at the scene, designs a sample learning model based on SVR-GA hybrid algorithm which completes two major tasks of knowledge acquisition and technical parameters selecting. The algorithm has been used to predict the flatness of the plate in the modeling industry and the result shows that the algorithm not only improves the Prediction accuracy but also has the ability to update in real-time online environment.