Selecting the forgetting factor in subset autoregressive modelling

Abstract. Conventional methods to determine the forgetting factors in autoregressive (AR) models are mostly based on arbitrary or personal choices. In this paper, we present two procedures which can be used to select the forgetting factor in subset AR modelling. The first procedure uses the bootstrap to determine the value of a fixed forgetting factor. The second procedure starts from this base and applies the time-recursive maximum likelihood estimation to a variable forgetting factor. In one illustration using real exchange rates, we demonstrate the effect of the forgetting factor in subset AR modelling on ex ante forecasting of non-stationary time series. In a second illustration, these two procedures are applied to time-update forecasts for a stock market index. Subset AR models not including a forgetting factor act as a set of benchmarks for assessing ex ante forecasting performance, and consistently improved forecasting performance is demonstrated for these proposed procedures.     

结合方向纹理熵的Haar-like特征在线boosting跟踪算法

针对传统基于Haar-like特征的on-line boosting跟踪算法（HBT）需要产生大规模随机特征、占用大量计算资源和存储空间的缺点,提出结合方向纹理熵的Haar-like特征在线boosting跟踪算法（HBTT）。HBTT算法利用灰度共生矩阵的熵获得目标纹理的方向信息,在此基础上有针对性地产生具有方向纹理信息的Haar-like特征,从而可有效避免无效随机特征的产生,减小特征池容量;更进一步,可根据目标纹理的复杂程度自动调整特征数量,使得算法更灵活。在跟踪过程中,在线学习模块可以使错误率较高的特征被结合了目标纹理方向信息的Haar-like特征所替换。与HBT算法比较,HBTT算法的跟踪误差降低了10%以上;在相同特征池容量下,置信度提高了2%以上。实验结果表明,该算法不仅具有较高的鲁棒性,而且在跟踪效率和性能上都有所提高。     

2,2′-oxo-1,1′-azobenzene: microbial transformation of rye (Secale cereale L.) allelochemical in field soils byAcinetobacter calcoaceticus: III

Acinetobacter calcoaceticus, a gram-negative bacterium isolated from field soil, was found to be responsible for the biotransformation of 2(3H)-benzoxazolinone (BOA) to 2,2-oxo-1,1-azobenzene (AZOB). Experiments were conducted to evaluate the transformation of BOA to AZOB by this microbe in sterile and nonsterile soil. Transformation studies with soils inoculated withA. calcoaceticus indicated that the production of AZOB increased linearly with the concentration of BOA in sterile soil and showed a quadratic trend in nonsterile soils. This also indicated that all soil types studied for the transformation experiments might containA. calcoaceticus capable of the conversion of benzoxazolinones.     

面向在线社区用户的群体推荐算法研究

结合现有两种主要群体推荐算法的优势, 建立新的算法框架, 并引入差异度因素对模型进行优化。另外, 考虑到在线社区用户的特点, 定义互动度指标来描述群体成员间的互动程度, 通过分析其与推荐精度之间的关系, 探讨互动度对群体推荐的影响。选取豆瓣网数据进行实验, 并与传统方法进行比较, 结果表明, 融入差异度的算法具有更好的推荐效果, 且有效的互动机制能够保证较高的推荐精度。     

考虑出行方向的广州出租车时空可达性研究

出行方向是可达性研究中的一个关键要素, 因此, 提出一种考虑出行方向的出租车时空可达性分析方法。根据出租车经验路径分方向构建经验等级路网, 采用基于分层路网的Dijkstra算法计算得到各OD对之间的行程时间, 作为可达性评价指标, 并采用网格法和反距离加权插值法绘制得到广州市中心商业区等时线及其渲染图, 分方向、分时段对其时空可达性进行分析。结果证明了不同出行方向下, 特别是处于高峰时段, 可达性有较大的不同。     

FINE CONTROL OF HZSM-5’S PORE-OPENING SIZE BY A NEW CVD METHOD

A new CVD method without vacuum condition(CVDWV)was designed in this study,ithad been found that by modifying the flow rate of the carrier gas(N_2)and the temperature of sur-face reaction,silicon saturator and sample tube,the amount of silica deposited on the external sur-face of the zeolite could be precisely controlled,The changes in the physicochemical properties ofzeolite,such as pore-opening size,acidity and reaction property were investigated.The results ob-tained showed that:modification of the improved CVD method did not change the internal structureand acidity of the zeolite,but could bring about significant change of pore-opening size and reactionproperty as desired.     

钛酸钾品须填充UHMWPE复合材料的摩擦磨损性能研究

用钛酸钾晶须(PTW)对超高摩尔质量聚乙烯(UHMWPE)进行填充改性,考察了复合材料的摩擦磨损性能,研究了其结晶情况,观察了磨损表面形貌并分析了其机理.结果表明随着PTW用量的增加,复合材料的硬度、结晶度以及维卡软化点都有所增大.PTW提高了复合材料的耐磨性,但摩擦系数有所上升.PTW的加入使得UHMWPE的磨损机理从黏着磨损和塑性变形改变成疲劳磨损和轻微的塑性变形,提高了其抗磨粒磨损的性能.     

内容分发网络中基于相关内容吸引的缓存算法

内容分发网络中基于内容名的缓存算法会导致路由表规模随网络增长而膨胀,将严重影响网络路由效率和性能。针对该问题,提出一种基于相关内容吸引的节点缓存算法。利用本地缓存算法,通过节点已缓存内容对其他内容的吸引作用吸引主要特征内容,排斥具有次要特征内容,将缓存中不同特征内容的数量差异进行放大,使缓存内容表现出明显稳定的内容特征。同时设计相关内容生存时间相互增强的缓存策略,以减少路由通告信息量,提高内容分发网络的路由能力。实验结果表明,该算法在有效解决路由问题的同时,能增强缓存内容稳定性,提高路由可信度。     

结晶梯恩梯生产工艺的研究

讨论了结晶梯恩梯的制造工艺、六硝基芪用量、络合物的形成与稳定性，以及六硝基芪质量等影响结晶梯恩梯质量的因素。     

一种改进的基于支持向量机的多类分类方法

针对现有支持向量机多类分类算法在分类精度上的不足,提出一种改进的支持向量机决策树多类分类算法。为了最大限度地减少误差积累的影响,该算法利用投影向量的思想作为衡量类分离性的标准,由此构建非平衡决策树,并且在决策树节点处对正负样本选取不同的惩罚因子来处理不平衡数据集的影响,最后引入KNN算法与SVM共同识别数据集。通过在手写体数字识别数据集上的仿真实验,分析比较各种方法,表明该方法能有效提高分类精度。