Los Angeles, CA
March 31, 2009 to April 2, 2009
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/CSIE.2009.393
This paper analyzes the defects and reasons for using standard BP neural network algorithm in building quality prediction model of yarns and explores an improved BP neural network algorithm. By increasing the back-propagation error-feedback signals and applying sell-adaptive and adjusting learning rate, the research has reinforced the adjustment of network weights and prevented network entering saturated region too early. These methods can increase the convergent speed of network and improve system stability. The experiment has proved that the forecast result is of high accuracy which comes from the improved BP neural network algorithm, and the design of quality prediction model is reasonable.
Fan Xiu-Juan, Li Cheng-Guo, "The Research in Yarn Quality Prediction Model Based on an Improved BP Algorithm", CSIE, 2009, 2009 WRI World Congress on Computer Science and Information Engineering, CSIE, 2009 WRI World Congress on Computer Science and Information Engineering, CSIE 2009, pp. 167-172, doi:10.1109/CSIE.2009.393