By combining block-wise recursive PLS with finite memory method, a new adaptive algorithm was proposed to build adaptive soft-sensor.
针对基于批量数据的传统偏最小二乘(PLS)模型无法随生产过程的变化而更新的问题,提出基于块式递推PLS的限定记忆法。
Kanerva's algorithm was an elegant method to store a finite number of data points in a very immense potential memory space.
卡内尔瓦的算法是一种将有限数量的数据点储存进非常巨大的潜在的内存空间的绝妙方法。
Compared with the finite element method, the method proposed in this paper requires less memory, less original data to input into computer, and has higher calculating accuracy.
与有限单元法相比,本方法所需的计算机内存少。原始数据输入较简单,算例表明,计算精度高。
应用推荐