Journal of Jianghan University(Natural Science Edition) ›› 2014, Vol. 42 ›› Issue (2): 92-96.

Previous Articles    

Computer Lab Failure Rate Prediction Model Based on BP Neural Network

ZHOU Siwei1,ZENG Ting2   

  1. 1.Department of Computer Science,Zhixing College of Hubei University,Wuhan 430011,Hubei,China;2.Periodicals Press of Jianghan University,Wuhan 430056,Hubei,China
  • Online:2014-04-25 Published:2014-05-15

Abstract: As information technology penetrating in various subject areas,more and more courses require students to complete the relevant operations in the computer lab. With increasing of computer users,computer loss increase. For better maintainability of lab,takes the historical failure rate of 210 computers in 2005 in computer lab of computer department of Zhixing college of Hubei University as sample,uses JAVA language,BP network training computer model to predict the failure rate of the batch of computers,and then contrasts the historical data to find some errors,and then uses the method of increasing momentum term to improve the BP algorithm,the improved training samples predicted results are basically consistent with historical data.

Key words: BP neural network, JAVA, method of increasing momentum term, failure rate of lab computer

CLC Number: