Diversified Kernel Latent Variable Space and Multi-Objective Optimization for Selective Ensemble Learning-Based Soft Sensor

The improvement of data-driven soft sensor modeling methods and techniques for the industrial process has strongly promoted development intelligent industry. Among them, ensemble learning is an excellent framework. Accuracy diversity are two key factors that run through entire stage building learning-based sensor. Existing base model generating or pruning always consider separately, which limited high-performance but low-complexity sensors. To work out this issue, a selective method based on multi-kernel latent variable space evolutionary multi-objective optimization proposed, referred to as MOSE-MLV-VSPLS. This designs multiple enhancement mechanism in generation stage. Diversified input subspaces first constructed using maximum information coefficient bootstrapping random resampling subset. Then set models combine accuracy generated supervised under kernel function perturbations. Further, quantifiable parameters designed diversity, gray wolf algorithm used select maximize these important achieve effective at MOSE-MLV-VSPLS applied typical industry processes, experimental results show superior ensemble-based modeling.