Both manifold learning and semi-supervised learning have been widely investigated in the past few years. Some of manifold learning algorithms, which are based on the idea of local approximation, can be used to control the way of transmitting information between point clouds. We combine local approximation with the idea of preserving projections and weighted integration, and give a set of solutions to semi-supervised regression and manifold alignment. Finally, we validated the effectiveness of the presented schemes in th...
