基于偏曩小二来回归法和模糊隶属度函数,提出了一种模糊偏曩小二秉支持向量机。传统最小二乘支持向量机引入模糊加权系数后,可以根据训练样本点的情况调整折衷系数,有效地提高了最小二乘支持向量机的抗噪性能。同时利用偏最小二乘回归法,克服了求解线性回归方程中自变量向量间的多重相关性问题。利用sinc函数对该建模方法进行了测试,并进一步对铜转炉吹炼时间的预测问题进行了仿真研究。仿真结果表明,该建模方法具有预测准确、跟踪性能好的优点。
On the base of partial least square regression and fuzzy memberships, a fuzzy partial least square support vector machine was proposed. After fuzzy weighting parameters were introduced into tradition least square support vector machine, tradeoff parameter could be tuned according to training samples. This could effectively increase the noise immunity of least square support vector machine. And the multiple correlation problems of independent variable vectors in equation of linear regression were solved by partial least square regression. This modeling method was tested by sinc function firstly. Then, the prediction problem about copper converter blowing time was studied by simulation. The results of simulation show the modeling method has the merits such as accurate prediction and good tracking performance.