Multi-manifold discriminant local spline embedding
作者：
Highlights：
• To preserve the geometry structure of data, MDLSE reconstructs all the manifolds compatibly in the low-dimensional embedding.
• MDLSE characterizes the local manifold patches with not only the neighborhood space but also the feature space of data.
• MDLSE constructs intrinsically smooth manifolds by maximizing the smoothness of the thin plate splines on each manifold.
• MDLSE maps the marginal data of different manifolds as far as possible to separate the whole manifolds economically.
摘要
•To preserve the geometry structure of data, MDLSE reconstructs all the manifolds compatibly in the low-dimensional embedding.•MDLSE characterizes the local manifold patches with not only the neighborhood space but also the feature space of data.•MDLSE constructs intrinsically smooth manifolds by maximizing the smoothness of the thin plate splines on each manifold.•MDLSE maps the marginal data of different manifolds as far as possible to separate the whole manifolds economically.
论文关键词：Manifold learning,Dimension reduction,Classification,Thin plate spline,Multiple manifolds
论文评审过程：Received 24 March 2021, Revised 22 January 2022, Accepted 15 April 2022, Available online 18 April 2022, Version of Record 23 April 2022.
论文官网地址：https://doi.org/10.1016/j.patcog.2022.108714
