This paper aims at formalizing the concept of learning rules from multisource data in a cardiac monitoring context. Our method has been implemented and evaluated on learning from d...
Many kernel learning methods have to assume parametric forms for the target kernel functions, which significantly limits the capability of kernels in fitting diverse patterns. Som...
Manydata mining algorithms developed recently are based on inductive learning methods. Very few are based on similarity-based learning. However, similarity-based learning accrues ...
Advances in technology have enabled new approaches for sensing the environment and collecting data about the world. Once collected, sensor readings can be assembled into data stre...
Eric P. Kasten, Philip K. McKinley, Stuart H. Gage
Annotating learning material with metadata allows easy reusability by different learning/tutoring systems. Several metadata standards have been developed to represent learning obje...