支持张量机的研究及其应用

张量学习在机器学习领域中有很好的研究前景和研究价值。本项目以最优化理论与方法为指导，从向量空间的支持向量机出发，研究以输入为张量的分类问题、回归问题和顺序回归问题建立支持张量机的优化模型和算法。即建立支持张量机优化模型，包括线性情形和非线性情形，其中特别着重研究非线性情形中的核技术；并且针对上述优化模型设计快速有效的求解算法。此外，还将构建模型，应用于流行性感冒疫苗设计。通过研究中提出新的支持张量机优化模型和算法，提炼出优化理论问题。

Tensor learning has the great potential and the bright prospect. This project is concerned with classification, regression and ordinal regression, where the inputs are tensors instead of vectors in the traditional situation. The aim is to establish the models and algorithms of the corresponding support tensor machines. The models include both the linear case and the nonlinear case, emphasising the kernel strategy in the later one. The algorithms should be efficient and fast. In addition, a practical problem – the influenza vaccine design problem will be studied via the support tensor machine proposed and analysed in this project.