Dong-Mei Li

  • Citations Per Year
Learn More
The features of highway traffic flow are analyzed. The modeling scheme to predict highway traffic flow is proposed according to these features. The formulas for training wavelet neural network are deduced after the researches on learning process of wavelet neural network. The training algorithm and procedure for wavelet neural network are designed. The(More)
In city transit planning, line network optimization is a good measure with low cost and high efficiency. The researches on line network optimization are important in all cities. The optimization method for public traffic line network is studied according to the feature of line network optimization and adjustment problem. The cluster indexes for bus network(More)
  • Dong-Mei Li
  • 2005 International Conference on Machine Learning…
  • 2005
A method for identification of chaotic systems with large noise based on regularized feedforward neural networks is proposed. The regularization method can improve greatly the generalization performance of the feedforward networks. At various noise levels, we train feedforward networks with regularization parameter and clarify fundamental properties of(More)
The optimization method for public traffic line network is studied according to the feature of line network optimization problem. By using system science theory, seven objective functions and three constraints of line network optimization are established. An ideal solution decision method based on entropy weight and multi-objective programming is proposed.(More)
Index weighting is an important issue in multiple attribute decision making. The application of combination weighting approach can overcome the limitations of using subjective weighting or objective weighting method only. It will help to reflect the essential characteristics of the evaluated object better. The paper discusses the combination weighting(More)
This article proposes the concept lattice reduction based on the improved discernibility matrix using discernibility matrix of rough set, and with examples, proves that the improved reduction algorithm costs shorter time, makes reduction speed faster, and also saves storage space in comparison with the original discernibility matrix reduction algorithm.(More)
The implication and feature of input and output efficiency of science and technology are analyzed. The evaluation index system of input and output efficiency of science and technology is established by using a new feature selection method based on principal component analysis. The input and output efficiency of science and technology of each province is(More)
A new feature selection method is proposed to establish the evaluation index system for allocation efficiency of science and technology resources in large-medium industrial enterprises. The allocation efficiency for each province is computed and evaluated by using data envelopment analysis method. The status of allocation efficiency in large- medium(More)
In this paper, we present that noisy chaotic systems can be identified with RBF neural networks. We design three-layers RBF network structure and clarify fundamental properties of RBF networks to learn noisy chaotic systems by some numerical experiments. We also evaluate the identified models with reconstruction of attractors by the identified models.(More)
A neural network control algorithm based on predictive control is proposed to control time-delay chaotic systems. When the time-delay chaotic model is unknown, the control system stabilizes a chaotic orbit onto an unstable fixed point without using the knowledge of the location of the unstable fixed point and the local linearized dynamics at the point. The(More)