Xu Qingyang

  • Citations Per Year
Learn More
K nearest neighbor classification algorithm (KNN) is one of relatively simple method in data mining classification techniques, study and research KNN classification algorithm for data classification and data mining technology has a very important significance. An attribute Weighted KNN(W-KNN) method is utilized for reducing the irrelevant attributes. And(More)
A modified radial basis function neural networks (RBFNN) model is proposed to solve the control problem that the flue temperature in coke oven usually has the properties of high nonlinearity, large time-delay and multiple disturbances. The proposed method adopts K-means to initialize hidden layer and center parameters of the network. Finally, the production(More)
  • Xu Qingyang
  • Proceedings of the 33rd Chinese Control…
  • 2014
Unmanned Surface Vehicle (USV) has been developed so fast due to the numerous applications. Course-keeping controller plays an important role in the navigation. PID is always adopted as the control law. However, tuning of PID is very difficult due to its multiple objectives. In this paper, NSGA-II is adopted to optimize the parameters of PID. In order to(More)
Based on the spectral data from SDSS, Kernel Support Vector Machines (K-SVM) is applied to classify quasars from other celestial body. Firstly, the basic theory of the SVM(Support Vector Machine) with relaxation factor and kernel function is introduced. Then, the main parameters are designed and selected. Finally, the method is applied to the classification(More)
  • 1