AbstractIn a wind tunnel, the Mach number in the test section is an important parameter that should be predicted quickly and accurately. In building a Mach number prediction model, large-scale and high-dimensional data is the main issue. To solve the issue, the feature subsets ensemble (FSE) method has been proposed. However, a major drawback of the FSE method is that a large number of submodels are necessarily combined. In this paper, the maximum entropy pruning (MEP) method is proposed to overcome this drawback in the FSE Mach number prediction model. The MEP method refers to finding a subset of submodels that best approximates the entire submodels, while maximizing the quadratic Rényi entropy criterion. Experiments demonstrate that, with much fewer submodels than the FSE and other Mach number models, the MEP-FSE Mach number model can improve the prediction performance (i.e., the generalization), and meet the requirements of the forecasting speed and the root mean square error (less than 0.002).
Wind Tunnel Mach Number Prediction Based on the Pruned Feature Subsets Ensemble Method
Journal of aerospace engineering ; 30 , 4
2017
Aufsatz (Zeitschrift)
Englisch
Wind Tunnel Mach Number Prediction Based on the Pruned Feature Subsets Ensemble Method
Online Contents | 2017
|The Mach number prediction based on the feature subsets ensemble in transonic wind tunnel
SAGE Publications | 2016
|The Mach number prediction based on the feature subsets ensemble in transonic wind tunnel
Online Contents | 2016
|