.png){kind=icon}
返回
The centered alignment multikernel nonhomogeneous gray model and its application
.png){kind=icon}
PRE
AI
DOI：10.1007/s11071-022-08019-9.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
The gray model is very useful for small sample prediction. Compared with the homogeneous gray model, the nonhomogeneous gray model (NGM) improves the prediction differential equation; thus, the NGM is able to effectively predict a nonhomogeneous exponential sequence. However, the NGM cannot be applied to nonlinear samples because of the linear whitening equation. Inspired by multikernel learning, the central alignment multikernel learning method is integrated into the NGM, and a centered alignment multikernel NGM (CAMKNGM) is presented in this paper. In our model, the kernel trick and kernel function are used to overcome the difficulties in solving the nonlinear mapping problems of the NGM. Additionally, more than one kernel function is utilized to generate responding kernel matrices by mapping the nonlinear features to different kernel spaces. Centered alignment multikernel learning optimization can be carried out according to the obtained kernel matrices to form a more sufficient feature space so that the combined kernel matrix can be used to greatly capture the nonlinear factors of the time series. Through multikernel learning optimization, the proposed model with unknown nonlinear transformation can be solved easily. In the experiment, we assess the performance of our model on three samples and compare the CAMKNGM with the single-kernel NGM (KRNGM) and other state-of-the-art models. The results show that the CAMKNGM model is obviously better than the support vector regression, NGM, optimized NGM and KRNGM.
Keyword:
Gray model
Kernel method
Learning kernels
NGM model
期刊
IF:
6
论文数:
1.4W
被引数:
4.1W
机构
引用论文
Forecasting short-term renewable energy consumption of China using a novel fractional nonlinear grey Bernoulli model基于新型分数阶非线性灰色伯努利模型的中国短期可再生能源消费预测.png)
RENEWABLE ENERGY
RENEWABLE ENERGYIF9.1
PREAI
分享收藏Trisomy 12 and elevated GLI1 and PTCH1 transcript levels are biomarkers for Hedgehog-inhibitor responsiveness in CLLBlood
OAAI
BloodIF0
OAAI分享收藏
查看更多内容