Extended Coopetitive Soft Gating Ensemble

摘要: This article is about an extension of a recent ensemble method called Coopetitive Soft Gating Ensemble (CSGE) and its application on power forecasting as well motion primitive cyclists. The CSGE has been used successfully in the field wind forecasting, outperforming common algorithms this domain. principal idea to weight models regarding their observed performance during training different aspects. Several extensions are proposed original within article, making even more flexible powerful. extended (XCSGE we term it), predict generation both wind- solar farms. Moreover, XCSGE applied forecast movement state cyclists context driver assistance systems. Both domains have requirements, non-trivial problems, evaluate various facets novel XCSGE. two problems differ fundamentally size data sets number features. Power based weather forecasts that subject fluctuations In cyclists, time delays contribute difficulty prediction. reaches improvement prediction up 11% for 30% compared worst performing model. For classification primitives 28%. evaluation includes comparison with other state-of-the-art methods. We can verify results significantly better using Nemenyi post-hoc test.