Manifold integration

From Wikipedia, the free encyclopedia
Jump to: navigation, search

Manifold integration is a combined concept of manifold learning and data integration, or an extension of manifold learning for multiple measurements.

Various manifold learning methods have been developed. However, they consider only one dissimilarity matrix corresponding to one kernel matrix, which represents one manifold of the data set. In practice, however, multiple sensors are used at a time, and each sensor generates data set on one manifold. In such a case, manifold integration is a desirable task, combining these dissimilarity matrices into a compromise matrix that faithfully reflects multiple sensory information on one integrated manifold.

For more information, see [1]

Notes[edit]

  1. ^ H. Choi, S. Choi and Y. Choe, "Manifold Integration with Markov Random Walks," in Proc. 23rd Association for the Advancement of Artificial Intelligence (AAAI-08), Chicago, Illinois, July 13–17, 2008.