Component Content Soft-Sensor in Rare-Earth Extraction Based on PSO and LS-SVM

To tackle the problem that the component content is difficult to detect online, an online prediction method of component content in the rare-earth extraction process using soft sensors based on least squares support vector machines (LS-SVM) is proposed. Particle swarm optimization algorithm (PSO) is presented to select the parameters of LS-SVM and the kernel function. The result of simulation indicates that this method is effective. Compared with the method base on neural network, the method based on LS-SVM is more effective to realize online prediction of the component content in the rare earth extraction process.