基于mRMR与因子分解机的分类模型研究

很多学者用“全球恐怖主义研究数据库”GTD数据集，采用博弈论、K近邻法和支持向量机等分析恐怖事件的聚集性，已经取得一些成果.但在前期研究中未有很好考虑数据的稀疏性以及高维度多冗余等会导致聚集分类准确率不高的问题.本文提出一种基于最小冗余最大相关与因子分解机结合的TFM分类模型，使用增量搜索方法寻找近似最优的特征解决高维度多冗余问题和FM方法解决数据稀疏问题，并对预处理后的恐怖袭击事件数据用TFM模型做量化分类.文中使用朴素贝叶斯NB、支持向量机SVM、逻辑回归LR与TFM等4个模型的“马修斯相关系数”MCC进行比较，结果显示TFM的MCC相对于其他三个模型NB、SVM、LR分别提高了49.9%,2.5%,2.3%，可见TFM模型有一定可行性.     

城乡结合部道路交叉口多目标信号配时优化模型

随着我国城镇化进程的加快,城乡结合部道路交通问题日益显著。为提高城乡结合部道路交叉口的交通运行效率、环境效益和交通安全,提出信号配时多目标优化方法,综合考虑延误、通行能力、停车次数、车辆尾气排放等因素构建模型,运用遗传算法进行求解。以某典型交叉口为例进行分析,给出改进设计方案并进行交通仿真评价。研究结果表明,通过多目标信号配时优化可以缓解城乡结合部道路交叉口交通拥堵和污染物排放的问题,具有较好的交通信号控制效果。     

基于频繁项集分类统计的增量式关联规则应用

针对商业交易数据构成项目繁多、动态数据增加量大、历史数据量更大的特点,根据频繁项集的商业特征,分为新生、成熟、老化、过期4种类型并分类统计;提出了基于分类统计增量地挖掘新增业务数据中关联规则的算法,算法只需两次扫描新增数据库,无需扫描历史数据库,算法将发现的规则按照其反应的商业特征分为4种类型:新生规则、成熟规则、老化规则、过期规则,在提升规则内容识别效率的同时,强化规则特点的识别能力。     

Role of humidity on the structures and properties of regenerated silk fibers

Silk fiber was processed from highly concentrated spinning dope to solid fibers along with water removal. To understand the mechanism of water removal during silk fiber spinning process, a microfluidic chip was designed and applied to investigate the structures and mechanical properties of two kinds of regenerated silk fibroin fibers dry-spun at different relative humidity. The experimental results showed that the diameters of the fibers spun at 40% RH are always larger than the fibers spun at 50% RH due to different removal rates of water. The fibers spun at low humidity contain more β-sheet structure and lower degree of chain orientation and crystalline orientation. These results indicate that the fast phase transition of silk fibroin from sol–gel to silk fiber undergoes with rapid water removal and higher fiber orientation relates to more residue water and drawing force.     

Efficient Feature Extraction Using Apache Spark for Network Behavior Anomaly Detection

Extracting and analyzing network traffic feature is fundamental in the design and implementation of network behavior anomaly detection methods. The traditional network traffic feature method focuses on the statistical features of traffic volume. However, this approach is not sufficient to reflect the communication pattern features. A different approach is required to detect anomalous behaviors that do not exhibit traffic volume changes,such as low-intensity anomalous behaviors caused by Denial of Service/Distributed Denial of Service(Do S/DDo S)attacks, Internet worms and scanning, and Bot Nets. We propose an efficient traffic feature extraction architecture based on our proposed approach, which combines the benefit of traffic volume features and network communication pattern features. This method can detect low-intensity anomalous network behaviors and conventional traffic volume anomalies. We implemented our approach on Spark Streaming and validated our feature set using labelled real-world dataset collected from the Sichuan University campus network. Our results demonstrate that the traffic feature extraction approach is efficient in detecting both traffic variations and communication structure changes.Based on our evaluation of the MIT-DRAPA dataset, the same detection approach utilizes traffic volume features with detection precision of 82.3% and communication pattern features with detection precision of 89.9%. Our proposed feature set improves precision by 94%.     

Bonding characteristics of the Al2O3-metal composite coating fabricated onto carbon steel by combustion synthesis

The fabrication of an alumina-metal composite coating onto a carbon steel substrate by using a self-propagating high-temperature synthesis technique was demonstrated. The effects of the type and thickness of the pre-coated layer on the binding structure and surface quality of the coating were systematically investigated. The macrostructure, phase composition, and bonding interface between the coating and the substrate were investigated by scanning electronic microscopy (SEM), X-ray diffraction (XRD), and energy-dispersive X-ray spectrometry (EDS). The diffraction patterns indicated that the coating essentially consisted of α-Al₂O₃, Fe(Cr), and FeO·Al₂O₃. With an increase in the thickness of the pre-coated working layer, the coating became more smooth and compact. The transition layer played an important role in enhancing the binding between the coating and the substrate. When the pre-coated working layer was 10 mm and the pre-coated transition layer was 1 mm, a compact structure and metallurgical bonding with the substrate were obtained. Thermal shock test results indicated that the ceramic coating exhibited good thermal shock resistance when the sample was rapidly quenched from 800°C to room temperature by plunging into water.     

一类含有2n个非零元的极小谱任意符号模式

设A为n阶符号模式,如果对任意n次首1实系数多项式r(x),都有一个实矩阵B在符号模式A的定性矩阵类Q(Α)中,且B的特征多项式为f(x)=r(x),则称A是谱任意的.如果A是谱任意的并且A的真子模式都不是谱任意的,则称A为极小谱任意的.文章对一类新的含有2n个非零元的n阶符号模式运用幂零——雅可比方法证明了其为极小谱任意模式.     

基于蚁群优化的命名数据网络QoS路由研究

针对命名数据网络(NDN)具有多样的路由转发策略,但都未能实现NDN的服务质量(QoS)保障能力,利用蚁群优化算法(ACO)设计相应的QoS路由机制(ACO_QoS)能够保障NDN的服务质量.实验结果表明:ACO_QoS能够在满足QoS的前提下找到一条花费最小的路由.