Although the twin supportvectorregression (TSVR) method has been widely studied and various variants are successfully developed, the structural risk minimization (SRM) principle and model's sparseness are not gi...
详细信息
Although the twin supportvectorregression (TSVR) method has been widely studied and various variants are successfully developed, the structural risk minimization (SRM) principle and model's sparseness are not given sufficient consideration. In this paper, a novel nonparallel support vector regression (NPSVR) is proposed in spirit of nonparallelsupportvector machine (NPSVM), which outperforms existing twin supportvectorregression (TSVR) methods in the following terms: (1) For each primal problem, a regularized term is added by rigidly following the SRM principle so that the kernel trick can be applied directly to the dual problems for the nonlinear case without considering an extra kernel-generated surface;(2) An epsilon-insensitive loss function is adopted to remain inherent sparseness as the standard supportvectorregression (SVR);(3) The dual problems have the same formulation with that of the standard SVR, so computing inverse matrix is well avoided and a sequential minimization optimization (SMO)-type solver is exclusively designed to accelerate the training for large-scale datasets;(4) The primal problems can approximately degenerate to those of the existing TSVRs if corresponding parameters are appropriately chosen. Numerical experiments on diverse datasets have verified the effectiveness of our proposed NPSVR in sparseness, generalization ability and scalability. (c) 2018 Elsevier Ltd. All rights reserved.
Basic oxygen furnace (BOF) steelmaking plays an important role in steelmaking process. Hence, it is necessary to study BOF steelmaking modeling. In this paper, a novel regression algorithm is proposed by using nonpara...
详细信息
Basic oxygen furnace (BOF) steelmaking plays an important role in steelmaking process. Hence, it is necessary to study BOF steelmaking modeling. In this paper, a novel regression algorithm is proposed by using nonparallel support vector regression with weight information (WNPSVR) for the end-point prediction of BOF steelmaking. The weight information is excavated by K-nearest neighbors (KNNs) algorithm. Since the whale optimization algorithm (WOA) has the characteristics of fast convergence speed and a few adjustment parameters, WOA is applied to optimize the parameters in the objective function of WNPSVR. Compared with traditional prediction models, WNPSVR-WOA is not easy to fall into local minimum values and is insensitive to noise. Thus, the prediction and control of molten steel end-point information are more accurate. Experimental results verify the effectiveness and feasibility of the proposed model. Within different error bounds (0.005 wt.% for carbon content model and 10 degrees C for temperature model), the hit rates of carbon content and temperature are 89% and 95%, respectively. Meanwhile, a double hit rate of 85% is achieved. The above results conclude that our WNPSVR-WOA has important reference value for actual BOF application and can improve the steel product quality. Moreover, WNPSVR-WOA can also be used to other fields.
Basic oxygen furnace(BOF) steelmaking plays a significant role in steelmaking ***,it is necessary to study the modeling of BOF *** order to realize the end-point prediction of converter steelmaking,improve the yield o...
详细信息
ISBN:
(纸本)9781665431293
Basic oxygen furnace(BOF) steelmaking plays a significant role in steelmaking ***,it is necessary to study the modeling of BOF *** order to realize the end-point prediction of converter steelmaking,improve the yield of target product and realize energy saving and emission reduction,an improved nonparallel support vector regression(INPS VR) algorithm is proposed in this ***,in order to speed up the modeling,whale optimization algorithm(WOA) is used to optimize the parameters of INPSVR *** has some guiding significance for small and medium converter enterprises to ensure tapping quality,improve production efficiency and reduce *** results show that the proposed prediction model has perfect performance in accuracy and efficiency.
暂无评论