Extreme learning machine (ELM) is an emergent method for training single hidden layer feedforward neural networks (SLFNs) with extremely fast training speed, easy implementation and good generalization performance. This work presents effective ensemble procedures for combining ELMs by exploiting diversity. A large number of ELMs are initially trained in three different scenarios: the original feature input space, the obtained feature subset by forward selection and different random subsets of features. The best combination of ELMs is constructed according to an exact ranking of the trained models and the useless networks are discarded. The experimental results on several regression problems show that robust ensemble approaches that exploit diversity can effectively improve the performance compared with the standard ELM algorithm and other recent ELM extensions.
Bones were obtained from three fish species (brown trout Salmo trutta m. fario, grayling Thymallus thymallus and Carpathian sculpin Cottus poecilopus) for regression analysis. Bones used were chosen based upon frequency of occurrence in spraint samples and diagnostic value. Relationships between the length of diagnostic bones and fish length, fish length and weight, and standard length to total length, were assessed for the three fish species. Polynomial regression was deemed most suitable for the relationship between bone length and fish standard length, multiplicative between fish standard length and fish weight, and linear (brown trout) or polynomial (grayling and Carpathian sculpin) for standard length against total length. All calculated regressions were highly significant and displayed high coefficients of determination, ranging between 93.9 and 99.8 %. The uses of the bones examined, and the equations produced, are discussed in the light of their future use in estimating prey numbers, length and biomass in otter diet analysis.
This article presents an application of evolutionary fuzzy rules to the modeling and prediction of power output of a real-world Photovoltaic Power Plant (PVPP). The method is compared to artificial neural networks and support vector regression that were also used to build predictors in order to analyse a time-series like data describing the production of the PVPP. The models of the PVPP are created using different supervised machine learning methods in order to forecast the short-term output of the power plant and compare the accuracy of the prediction.