Application Research Of MLSSVR Based On IGSA

Optimization has received widespread attention as an important scientific branch. It has influenced many disciplines greatly and obtained rapid promotion and application in so many engineering domains. Optimization has become an indispensable tool for many different areas. This paper put forward Improved Gravitational Search Algorithm(IGSA) based on the standard Gravitational Search Algorithm(GSA) and designed flatness pattern recognition of Multi-output Least Squares Support Vector Regression(MLSSVR) model based on IGSA. According to the above content, this article has carried on the work as follows.First of all, in order to overcome the defects of GSA at the same time to improve the development and exploration ability, this paper proposed IGSA, which referenced the flight characteristics of wild goose team and method of weighted mean, Benchmark functions were used to verify the optimization effection of the improved algorithm and the results showed that IGSA not only had better convergence accuracy and faster convergence speed, but also searched more effectively in global.Second, the standard Least Squares Support Vector Regression(LS-SVR) just applied to single-output system, but industrial processes were multi-output system mostly. MLSSVR model has great potential for application. IGSA-MLSSVR model was designed which IGSA was used to optimize parameters, and was applied to the flatness pattern recognition. The simulation results show that the model has not only high recognition accuracy, but also strong generalization ability.Finally, for the purpose of displaying the recognition effect more conveniently and intuitively, using GUI interface in MATLAB environment to achieve the construction of flatness pattern recognition simulation system visual virtual platform based on IGSA-MLSSVR. It contains the simulation model and its simulation results and provides an efficient solution for flatness recognition simulation system.