基于遗传二进制粒子群混合算法的测试点决策研究

测试点选择是测试性分析和设计的前提。通过对测试问题的分析,建立了数学模型和构造了衡量测试集优劣程度的启发式函数,提出了一种遗传-二进制粒子群混合算法求解满足测试性指标要求的最小完备子集。应用实例表明,该算法能够有效的克服单一算法陷入局部最优和早熟收敛等不足,提高了搜索效率,能够有效快速的获得全局最优解。     

基于代价敏感支持向量机的推荐系统托攻击检测方法

基于标准支持向量机的托攻击检测方法不能体现由于用户误分代价不同对分类效果带来的影响,提出了一种基于代价敏感支持向量机的托攻击检测新方法,该方法在代价敏感性学习机制下引入支持向量机作为分类工具,对支持向量机输出进行后验概率建模,建立了基于类别隶属度的动态代价函数,更准确地反映不同样本的分类代价,在此基础上设计了代价敏感支持向量机分类器。将该分类器应用在推荐系统托攻击检测中,并与标准的支持向量机方法、代价敏感支持向量机方法进行比较,实验结果表明,本方法可以更精确地控制代价敏感性,进一步提高对攻击用户的检测精度,降低总体的误分类代价。     

支持跨步访问的嵌入式存储系统

提出了一种新型多素数嵌入式存储系统,能够显著改善系统跨步访问的性能。提高跨步访存的带宽,对于改善系统的整体性能有着重要的意义。但是,在嵌入式系统中,受片外结构的尺寸限制,直接应用经典的素数存储系统理论无法显著改善跨步访存性能。为此,该新型系统以素数存储系统理论为基础,引入主存访问调度策略并结合嵌入式系统的实际结构特征,构造了一种两层结构的多素数存储系统,可以用较少数量的存储模块实现,而且从逻辑地址到物理地址的映像计算简单,能够以相对较小的硬件代价实现对嵌入式存储系统跨步访问的有效支持。理论分析和实验结果均证实了该系统的正确性和有效性。     

Microstructure and Properties of Co2+: ZnAl2O4/SiO2 Nanocomposite Glasses Prepared by Sol–Gel Method

Transparent bulk Co²⁺: ZnAl₂O₄/SiO₂ nanocomposites containing nanocrystalline Co²⁺: ZnAl₂O₄ dispersed in silica glass matrix were obtained by the sol–gel method. The gels of composition 89SiO₂–6Al₂O₃–5ZnO− x CoO ( x =0.2, 0.4, 0.6, 0.8, 1.0) (mol%) were prepared at room temperature by using two different aluminum salts, aluminum nitrate and aluminum alkoxide (aluminum-iso-propoxide, Al(OPrⁱ)₃), as starting materials. The transparent gels were converted to the crystalline phase of gahnite by heating above 900°C. The microstructural evolution of gels was characterized. The effect of Co²⁺ concentration on spectroscopic properties was also discussed. Co²⁺: ZnAl₂O₄ nanocrystals dispersed in the SiO₂-based glass are formed at lower heat-treatment temperature and shorter heating time by using Al(OPrⁱ)₃ as raw material.     

组合KdV方程的Hamilton系统

本文根据KdV方程的Hamilton系统,构造并证明了组合KdV方程的Hamilton系统。     

基于通道注意力机制和U-net的医学图像分割方法

随着人工智能和医学大数据的发展,基于深度学习的医学图像分割技术因具有重要的应用价值和前景,已经成为目前的研究热点.为了增强特征图的语义信息,在U-net网络的基础上引入通道注意力机制,对U-net生成的特征逐通道进行压缩,将压缩后的特征逐通道计算权重,然后将该权重与原始特征相乘得出最终的特征.通过在两个不同器官的医学图像数据集上进行实验,Dice系数相较于原始U-net网络分别提高了2.7％和1.8％,验证了该方法的可行性和有效性.     

Augmented Online Sequential Quaternion Extreme Learning Machine

Neural Processing Letters - Online sequential extreme learning machine (OS-ELM) is one of the most popular real-time learning strategy for feedforward neural networks with single hidden layer due...     

Visualizing and understanding graph convolutional network

Multimedia Tools and Applications - The graph convolutional network (GCN), which can handle graph-structured data, is enjoying great interest in recent years. However, while GCN achieved remarkable...     

LS—SVM算法在光伏短期功率预测中的应用

以光伏阵列为研究对象,分析了辐照强度、温度以及日类型对光伏阵列出力的影响。建立了光伏短期功率预测最小二乘支持向量机LS-SVM模型。依据实验数据对模型进行了验证计算．并与BP神经网络模型做了比较,其中LS-SVM模型最大相对误差值为10．54％,平均绝对百分比误差（MAPE）为8．18％,绝对误差平方和平均值的均方根（RMSE）为0．4884,表明模型预测值离散化程度较小,所有预测点均与实际值非常接近,模型具有较好的拟合效果和泛化能力．可以有效地预测短期光伏发电功率。     

Degradability and biocompatibility of bioglass/poly(amino acid) composites with different surface bioactivity as bone repair materials

Bioglass (BG) possesses excellent bioactivity and has been widely used in the manufacture of biomaterials. In this study, a composite with different surface bioactivity was fabricated via in situ melting polymerization by incorporating BG and poly(amino acid) (PAA) at a suitable ratio. The structure of the composite was characterized by Fourier transform infrared spectroscopy and XRD. The compressive strength of the BG/PAA composites was 139 MPa (BG:PAA = 30:70). The BG/PAA composites were degradable, and higher BG in composite showed higher weight loss after 4 weeks of incubation in simulated body fluid. In addition, the BG/PAA composite maintained adequate residual compressive strength during the degradation period. The SEM results showed the differences in surface bioactivities of the composites directly, and 30BG/PAA composite showed thicker apatite layer and higher Ca/p than 15BG/PAA. in vitro MG-63 cell culture experiments showed that the composite was noncytotoxic and thus allows cells to adhere, proliferate, and differentiate. This indicates that the composite has good biocompatibility. The implantations in the bone defects of rabbits for 4 and 12 weeks were studied. The composites had good biocompatibility and were capable of guiding new bone formation without causing any inflammation. The composite may be successfully used in the development of bone implants.