Wang L.,University of Science and Technology Beijing |
Wang L.,Key Laboratory of Advanced Control of Iron and Steel Process Ministry of Education |
Zhang D.,Beijing Key Laboratory of Materials Science Knowledge Engineering |
Wu L.,University of Science and Technology Beijing |
Wu L.,Key Laboratory of Advanced Control of Iron and Steel Process Ministry of Education
Journal of Computational Information Systems | Year: 2013
Regression estimation is to model a desired function or an input-output relation from a set of input-output sample data. In this paper we present a fuzzy model construction approach to extract well-defined and semantically sound information granules from numerical data based on a multi-level clustering framework which performs three main clustering steps on the data space in order to extract granules qualitatively described in terms of fuzzy sets that meet interpretability constraints. Further, the TSK (Takagi-SugenoKang) fuzzy model is constructed with fuzzy if-then rules of which the SVR (support vector regression) method is integrated into the consequent part of the rule simultaneously to effectively perform the regression estimation problem. Two examples are illustrated and their results have shown the proposed approach has better performance in various kinds of data domains than the traditional SVR. © 2013 by Binary Information Press.
Zou L.-H.,University of Science and Technology Beijing |
Zou L.-H.,Beijing Key Laboratory of Materials Science Knowledge Engineering |
Zhang D.,University of Science and Technology Beijing |
Zhang D.,Beijing Key Laboratory of Materials Science Knowledge Engineering
Journal of Theoretical and Applied Information Technology | Year: 2012
Artificial intelligence (AI) is an interdisciplinary research and it has been widely spread into many specific domains. In this paper, data engineering, knowledge engineering, innovation methodology and intelligent application technology of AI are mainly presented and their prospects in materials science are summarized after analyzing the intelligent service demands in nowadays materials engineering. The work of this paper will be helpful for new materials development and design and has reference value to the future research directions in materials informatics. © 2005 - 2012 JATIT & LLS. All rights reserved.
Xie Y.,University of Science and Technology Beijing |
Xie Y.,Beijing Key Laboratory of Materials Science Knowledge Engineering |
Lu N.,University of Science and Technology Beijing |
Lu N.,Beijing Key Laboratory of Materials Science Knowledge Engineering |
And 6 more authors.
Journal of Information and Computational Science | Year: 2015
Rolling force is one of the most important parameters during the process of strip steel hot continuous rolling, since rolling force model is a factor that directly affects the controlling accuracy of strip steel's thickness. The more accurate the rolling force model is, the better the need for production development gets. In this paper, firstly, a hybrid algorithm of the Least Squares Support Vector Regression machine (LS-SVR) and Particle Swarm Optimization (PSO) algorithm is proposed. Then the parameters that influencing the rolling force model can be identified by utilizing measured data, which, got during finishing rolling process. Finally, these parameters have to be optimized and the optimal solution is obtained. By this means, we successfully solve the problem of the identification of rolling force model. Then we also solve the problems in actual production. © 2015 by Binary Information Press