ID 原文 译文
10664 仿真结果表明,该算法能够有效地提高分类精度,且在不同试验数据下具备良好的稳定性。 The simulation results show that the algorithm can effectively improve the classification accuracy, and good stability under different test data.
10665 提取稳定有效的目标特征对于低分辨雷达的目标识别分类有着重要意义。 Stable and effective target feature extracting for low resolution radar target recognition classification has important significance.
10666 在提取目标基本特征雷达散射截面积(radar cross section,RCS)与频谱熵值的基础上,提出了一种基于特征概率分布曲线的目标分类方法。 ‭Basic features in the extraction of target radar scattering cross section area (radar cross section, the RCS) and spectral entropy, on the basis of this paper proposes a classification method based on the characteristics of probability distribution curve of the target.
10667 该方法首先应用快速傅里叶变换计算目标回波频谱,提取目标RCS与频谱熵值。 ‭This method first fast Fourier transform are used to calculate the target echo spectrum and extract the target RCS and spectrum entropy.
10668 然后滑窗分段计算基本特征的概率分布曲线,从而利用概率分布曲线提取出稳定的目标特征,最后利用支持向量机对目标实现分类。 And then the basic features of sliding window section calculation probability distribution curves, which using the probability distribution curve to extract the target characteristics of stability, the final classification using support vector machine (SVM) to the target.
10669 基于实测数据的分类结果表明,该特征具有较好的稳健性和分类性能,同时算法便于工程实现。 ‭Based on the measured data of the classification results show that the feature has good robustness and classification performance, at the same time algorithm is convenient for engineering implementation. ‭
10670 针对区间灰数的灰色聚类模型中指标权重确定的问题,借鉴信息熵的思想,引入灰色熵权确定指标权重,构造了基于熵权和区间灰数信息的聚类评估算法。 For grey clustering model of interval grey number determination of index weight problem, draw lessons from the idea of information entropy, introducing grey entropy to determine index weight, is constructed based on information entropy and interval grey number of clustering evaluation algorithm.
10671 该算法以区间灰数本身的信息为依据通过计算灰熵来得到聚类指标权重。 With interval grey Numbers for the algorithm on the basis of the information itself is obtained by calculating grey entropy to the clustering index weight.
10672 最后以实际问题为背景进行算例研究,结果表明由所提算法所得的归一化聚类系数矩阵区分度更好,验证了所提算法的有效性和可行性。 Finally in the actual problem as the background for case study, the results indicate that the normalized obtained by the proposed algorithm better clustering coefficient matrix differentiation, verify the feasibility and effectiveness of the proposed algorithm.
10673 为了解决考虑不确定性时仿真系统动态输出验证问题,提出一种基于数据特征的仿真系统动态输出验证方法。 In order to solve dynamic output considering uncertainty simulation system validation issues, put forward a method based on dynamic output characteristics of the data simulation system verification.