RESUMEN
This paper presents a theoretical foundation for an SVM solver in Krein spaces. Up to now, all methods are based either on the matrix correction, or on non-convex minimization, or on feature-space embedding. Here we justify and evaluate a solution that uses the original (indefinite) similarity measure, in the original Krein space. This solution is the result of a stabilization procedure. We establish the correspondence between the stabilization problem (which has to be solved) and a classical SVM based on minimization (which is easy to solve). We provide simple equations to go from one to the other (in both directions). This link between stabilization and minimization problems is the key to obtain a solution in the original Krein space. Using KSVM, one can solve SVM with usually troublesome kernels (large negative eigenvalues or large numbers of negative eigenvalues). We show experiments showing that our algorithm KSVM outperforms all previously proposed approaches to deal with indefinite matrices in SVM-like kernel methods.
RESUMEN
BACKGROUND: Quality-controlled subject gateways are Internet services which apply a selected set of targeted measures to support systematic resource discovery. Considerable manual effort is used to process a selection of resources which meet quality criteria and to display a extensive description and indexing of these resources with standards-based metadata. OBJECTIVE: Several metadata element sets are proposed to describe, index and quality health resources to be included in a French quality-controlled health gateway called CISMeF. The main objectives were to enhance Internet health document retrieval and navigation, and to allow interoperability with other Internet services. RESULTS: The Dublin Core metadata element set is used to describe and index all Internet health resources included in CISMeF. For teaching resources, some elements from IEEE1484 Learning Object Metadata are also used. For evidenced-base medicine resources, specific metadata are employed which assess the health content quality. The HIDDEL metadata set is used to enhance transparency, trust and quality of health information on the Internet. CONCLUSION: Comprehensive metadata element sets can be extremely useful to describe, index and assess health resources on the Internet in a quality-controlled subject gateway. Machine-readable metadata creates an Semantic Web which is more efficient for end-users as compared to the current Web.