ID 原文 译文
4533 仿真结果表明，推导的符合度统计特性与实际情况一致，所提算法能在低信噪比下有效完成参数识别，同时具有较好的低信噪比适应能力，在信噪比为 5 dB，码长为 511 的条件下，能够完成参数的可靠识别， The simulation results show that the deduced statistical character-istics are consistent with the actual situation, and the proposed algorithm can achieve reliable recognition under SNR of 5 dBand code length of 511.
4534 与目前算法相比，所提算法性能优于现有的软判决算法，且比硬判决算法提升 1~3.5 dB。 Comparing with existing algorithms, the performance of the proposed algorithm is better than that of the existing soft-decision algorithm and 1~3.5 dB better than that of the hard-decision algorithms.
4535 针对传统前向后向匹配追踪（FBP）算法运行时间较长的问题，提出了一种自适应加速前向后向匹配追踪（AAFBP）算法。 Aiming at the long running time problem of the traditional forward-backward pursuit (FBP) algorithm, an adap-tive acceleration forward-backward pursuit (AAFBP) algorithm was proposed.
4536 AAFBP 算法的重构过程可分为 2 个阶段，在前向阶段利用自适应阈值来选取适量原子加入支撑集， The reconstruction process of AAFBP algo-rithm can be divided into two stages. In the forward stage, the AAFBP algorithm used the adaptive threshold to select the rightamount of atoms to join the support set.
4537 在后向回溯过程中以原子的投影系数大小作为删除依据，利用自适应删除阈值来进行原子的删除，同时克服了自适应过程中存在的回溯过度现象。 In the backward stage, based on the projection coefficient of the atoms, the deletionthreshold was introduced to remove the atoms adaptively and the excessive backtracking phenomenon in adaptive process wasovercome simultaneously.
4538 所提方法能够保证选入原子数量更具随机性，使每次迭代保留更多的正确原子。 The proposed method can ensure the number of the selected atoms more random, and more rightatoms were retained in each iteration.
4539 一维稀疏信号和二维图像的仿真结果表明，AAFBP 算法在重构精度和运算时间上都更具有优势。 The simulation results of one-dimensional sparse signal and two-dimensional image show that the AAFBP algorithm has more advantages in both the accuracy of reconstruction and the running time.
4540 针对入侵检测系统使用单个事件作为攻击检测的特征会导致较高误报率的问题，提出了利用贝叶斯网络模型进行跨空间的事件关联和利用卡尔曼滤波器线性模型进行跨时间的事件关联的内网攻击检测方法。 In view of the fact that a single event as an attack detection feature leads to a higher false positive rate, an in-tranet attack detection method using Bayesian network model for cross-space event correlation and Kalman filter linearmodel for cross-temporal event correlation was proposed.
4541 基于该方法实现了一个进程查询系统，该系统可以根据用户的高层过程描述来扫描和关联分布的网络事件。 Based on the method, a process query system was implemented,which can scan and correlate distributed network events according to the user's high-level process description.
4542 实验分析表明，该方法在不增加明显计算开销的情况下能够显著减少内网攻击检测的误报率。 Experi-mental analysis show that the proposed method can significantly reduce the false positive rate of intranet attack detectionwithout increasing the computational overhead.