Reliable forecasts of the price of oil are of interest for a wide range of applications. For example, central banks and private sector forecasters view the price of oil as one of the key variables in generating macroe...Reliable forecasts of the price of oil are of interest for a wide range of applications. For example, central banks and private sector forecasters view the price of oil as one of the key variables in generating macroeconomic projections and in assessing macroeconomic risks. Of particular interest is the question of the extent to which the price of oil is helpful in predicting recessions. This paper presents a statistical learning method （SLM） based on combined fuzzy system （FS）, artificial neural network （ANN）, and support vector regression （SVR） to cope with optimum long-term oil price forecasting in noisy, uncertain, and complex environments. A number of quantitative factors were discovered from this model and used as the input. For verification and testing, the West Texas Intermediate （WT1） crude oil spot price is used to test the effectiveness of the proposed learning methodology. Empirical results reveal that the proposed SLM-based forecasting can model the nonlinear relationship between the input variables and price very well. Furthermore, in-sample and out-of-sample prediction performance also demonstrates that the proposed SLM model can produce more accurate prediction results than other nonlinear models.展开更多
Prey choice is often evaluated at the species or population level. Here, we analyzed the diet of octopuses of different populations with the aim to assess the importance of individual feeding habits as a factor affect...Prey choice is often evaluated at the species or population level. Here, we analyzed the diet of octopuses of different populations with the aim to assess the importance of individual feeding habits as a factor affecting prey choice. Two methods were used, an assessment of the extent to which an individual octopus made choices of species representative of those population （PSi and IS） and 25% cutoff values for number of choices and percentage intake of individual on their prey. In one population of Octopus cfvulgaris in Bermuda individuals were generalist by IS=0.77, but most chose many prey of the same species, and were specialists on it by 〉75% intake. Another population had a wider prey selection, still generalist with PSi=0.66, but two individuals specialized by choices. In Bonaire, there was a wide range of prey species chosen, and the population was specialists by IS=0.42. Individual choices revealed seven specialists and four generalists. A population of Octopus cyanea in Hawaii all had similar choices of crustaceans, so the population was generalist by IS with 0.74. But by individual choices, three were considered a spe-cialist. A population of Enteroctopus dofleini from Puget Sound had a wide range of preferences, in which seven were also spe-cialists, IS=0.53, By individual choices, thirteen were also specialists. Given the octopus specialty of learning during foraging, we hypothesize that both localized prey availability and individual personality differences could influence the exploration for prey and this translates into different prey choices across individuals and populations showed in this study.展开更多
In this paper,a decentralized iterative learning control strategy is embedded into theprocedure of hierarchical steady-state optimization for a class of linear large-scale industrial processeswhich consists of a numbe...In this paper,a decentralized iterative learning control strategy is embedded into theprocedure of hierarchical steady-state optimization for a class of linear large-scale industrial processeswhich consists of a number of subsystems.The task of the learning controller for each subsystem is toiteratively generate a sequence of upgraded control inputs to take responsibilities of a sequential stepfunctional control signals with distinct scales which are determined by the local decision-making units inthe two-layer hierarchical steady-state optimization processing.The objective of the designated strategyis to consecutively improve the transient performance of the system.By means of the generalized Younginequality of convolution integral,the convergence of the learning algorithm is analyzed in the sense ofLebesgue-p norm.It is shown that the inherent feature of system such as the multi-dimensionality andthe interaction may influence the convergence of the non-repetitive learning rule.Numerical simulationsillustrate the effectiveness of the proposed control scheme and the validity of the conclusion.展开更多
文摘Reliable forecasts of the price of oil are of interest for a wide range of applications. For example, central banks and private sector forecasters view the price of oil as one of the key variables in generating macroeconomic projections and in assessing macroeconomic risks. Of particular interest is the question of the extent to which the price of oil is helpful in predicting recessions. This paper presents a statistical learning method （SLM） based on combined fuzzy system （FS）, artificial neural network （ANN）, and support vector regression （SVR） to cope with optimum long-term oil price forecasting in noisy, uncertain, and complex environments. A number of quantitative factors were discovered from this model and used as the input. For verification and testing, the West Texas Intermediate （WT1） crude oil spot price is used to test the effectiveness of the proposed learning methodology. Empirical results reveal that the proposed SLM-based forecasting can model the nonlinear relationship between the input variables and price very well. Furthermore, in-sample and out-of-sample prediction performance also demonstrates that the proposed SLM model can produce more accurate prediction results than other nonlinear models.
文摘Prey choice is often evaluated at the species or population level. Here, we analyzed the diet of octopuses of different populations with the aim to assess the importance of individual feeding habits as a factor affecting prey choice. Two methods were used, an assessment of the extent to which an individual octopus made choices of species representative of those population （PSi and IS） and 25% cutoff values for number of choices and percentage intake of individual on their prey. In one population of Octopus cfvulgaris in Bermuda individuals were generalist by IS=0.77, but most chose many prey of the same species, and were specialists on it by 〉75% intake. Another population had a wider prey selection, still generalist with PSi=0.66, but two individuals specialized by choices. In Bonaire, there was a wide range of prey species chosen, and the population was specialists by IS=0.42. Individual choices revealed seven specialists and four generalists. A population of Octopus cyanea in Hawaii all had similar choices of crustaceans, so the population was generalist by IS with 0.74. But by individual choices, three were considered a spe-cialist. A population of Enteroctopus dofleini from Puget Sound had a wide range of preferences, in which seven were also spe-cialists, IS=0.53, By individual choices, thirteen were also specialists. Given the octopus specialty of learning during foraging, we hypothesize that both localized prey availability and individual personality differences could influence the exploration for prey and this translates into different prey choices across individuals and populations showed in this study.
基金supported by the National Natural Science Foundation of China under Grant No. F030101 60574021.
文摘In this paper,a decentralized iterative learning control strategy is embedded into theprocedure of hierarchical steady-state optimization for a class of linear large-scale industrial processeswhich consists of a number of subsystems.The task of the learning controller for each subsystem is toiteratively generate a sequence of upgraded control inputs to take responsibilities of a sequential stepfunctional control signals with distinct scales which are determined by the local decision-making units inthe two-layer hierarchical steady-state optimization processing.The objective of the designated strategyis to consecutively improve the transient performance of the system.By means of the generalized Younginequality of convolution integral,the convergence of the learning algorithm is analyzed in the sense ofLebesgue-p norm.It is shown that the inherent feature of system such as the multi-dimensionality andthe interaction may influence the convergence of the non-repetitive learning rule.Numerical simulationsillustrate the effectiveness of the proposed control scheme and the validity of the conclusion.
