This paper studies the addition of linear constraints to Support Vector Regression when kernel is linear. Adding those into problem allows add prior knowledge on estimator obtained, such as finding positive vector, probability vector or monotone data. We prove that related optimization stays a semi-definite quadratic problem. also propose generalization Sequential Minimal Optimization algorithm...

In this paper, we applied the support vector machine (SVM) to the spatial interpolation of the multi-year average annual precipitation in the Three Gorges Region basin. By combining it with the inverse distance weighting and ordinary kriging method, we constructed the SVM residual inverse distance weighting, as well as the SVM residual kriging precipitation interpolation model and compared them...

The arrangement of plant species along elevational gradients is prominent in the debate between individualistic versus community-unit concepts ecology. We obtained ranges (upper and lower elevation limits) for woody ferns on three highest mountains Aotea (Great Barrier Island), Aotearoa-New Zealand. These data allowed potential composition to be any were analysed using regression, ordination, c...