基于CG-EGU神经网络的计算机软件性能评估研究  

Research on computer software performance evaluation based on CG-EGU neural network

作　　者：刘冰洲 刘微[1] LIU Bing-zhou;LIU Wei(The Fourth Affiliated Hospital of Harbin Medical University,Harbin 150000,China)

机构地区：[1]哈尔滨医科大学附属第四医院信息中心,哈尔滨150000

出　　处：《信息技术》2025年第3期163-169,共7页Information Technology

摘　　要：计算机领域的学者针对软件系统提出了上百种性能评估模型,但随着计算机技术的发展,软件系统性能评估的难度大大提升。为此,该研究提出了一种新的带控制门的高效门控单元神经网络计算机软件评估模型,在更新门的基础上引入重置门,提高模型学习训练的效率,其次在两个隐含层间加入控制门,提高模型信息学习的能力。实验证明,改进后模型准确率比改进前增加了近20%,平均误差值比改进前降低50%。综合各指标表明,该研究提出的改进模型对计算机软件性能的评估效果更好,能为软件评估领域提供科学合理的评估依据。Scholars in the computer field have proposed hundreds of performance evaluation models for software systems,but with the development of computer technology,the difficulty of software system performance evaluation has greatly increased.To this end,a new efficient gating unit neural network computer software evaluation model with control gates is proposed.A reset gate is introduced on the basis of the updating gate to improve the efficiency of model learning and training.Secondly,a control gate is added between the two hidden layers to improve the ability of model information learning.The experiment shows that the accuracy of the improved model has increased by nearly 20%,and the average error value has decreased by 50%.The comprehensive evaluation of various indicators indicates that the improved model proposed in the study has a better evaluation effect on computer software performance,and can provide a scientific and reasonable evaluation basis for the field of software evaluation.

关 键 词：深度学习 神经网络 CG-EGU 软件性能评估 准确率 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]