因素空间理论下基点分类算法研究

目前，基于因素空间理论的背景基提取算法计算过程复杂，初始化必须依赖各因素极值，基点数量提取冗余等原因，未能在应用中取得很好效果。为此，结合内点判别法和知识可继承、可扩展的思想，提出一种计算简单、初始化独立、基点数量小的改进的背景基提取算法。然后，利用改进的背景基提取算法构造出一种全新的数据分类算法-基点分类算法，基点分类算法以提取每一类样本的背景基为预测模型，再通过新定义的λ-背景基，优化预测模型。数值实验表明:基点分类算法原理简单、构造难度小、分类模型泛化能力强，预测能力准确率高，同时严格的模型限定区域又能为识别新类别提供新方法。     

基于归一化水体指数及其阈值自适应算法的水体遥感反演效果分析

归一化水体指数（NDWI）是水体遥感反演的一种重要指标，其阈值及修正直接影响反演结果的精度。基于Landsat-8与GF-1光学多光谱影像，使用归一化水体指数法阈值0（TH₀）、最大类间方差法（OTSU）自适应阈值（TH_otsu）与均值漂移聚类算法（Mean-Shift）自适应阈值（TH_MS）分别对典型正常水体、云雾覆盖水体、富营养化水体、高含泥沙水体进行水体遥感提取与效果分析，结果表明：正常水体以TH₀为阈值提取精度最高，THMS提取精度次之，TH_otsu提取精度最差；而云雾覆盖水体、富营养化水体以及含泥沙水体使用THMS提取精度最高，尤其少量云雾覆盖下的水体，TH_MS具有更明显的优势，TH₀提取精度次之，TH_otsu提取精度最差；对于不同的阈值，Land?sat-8比GF-1总体表现出更高的水体提取精度。Mean-Shift算法应用于NDWI阈值修正与水体遥感反演具有快速、水质适应性强、效果稳定的优势，对尤其是复杂条件下的水体信息遥感反演具有较好的提取效果。     

Air pollution and meteorological processes in the growing dryland city of Urumqi (Xinjiang, China)

Seven years (2000-2006) of monthly PM₁₀ (particulate matter, d ≤ 10 μm), SO₂, and NO₂ concentrations are reported for Urumqi, the capital of Xinjiang in NW China. Considerably high mean annual concentrations have been observed, which ranged between 150 and 240 μg m^− 3 (PM₁₀), 31 and 50 μg m^− 3 (NO₂), and 49 and 160 μg m^− 3 (SO₂). The shapes of seasonal variation of all pollutants were remarkably similar; however, winter/summer ratios of concentrations were quite different for PM₁₀ (2-3) and NO₂ (≈ 4) compared to SO₂ (up to 30). Very high consumption rates of fossil fuels for energy generation and domestic heating are mainly responsible for high annual pollution levels, as well as the (very) high winter/summer ratios. Detailed analysis of the 2000-2006 records of Urumqi's meteorological data resulted in inter-annual and seasonal frequency distributions of (a) (surface) inversion events, (b) heights of surface inversions, (c) stability classes of Urumqi's boundary layer, and (d) the “Air Stagnation Index (ASI)”. Urumqi's boundary layer is shown to be characterized by high mean annual and seasonal frequencies of (surface) inversions and by the dominance of stable dispersion classes. A further outcome of the meteorological analysis is the proof of Urumqi's strong diurnal wind system, which might have particularly contributed to the stabilization of the nocturnal boundary layer. Annual and seasonal variations of pollutant's concentrations are discussed in the context of occurrences of inversions, boundary layer, stability classes, and ASI. The trend of Urumqi's air pollution indicates a strong increase of mean annual concentrations 2000-2003, followed by a slight increase during 2003-2006. These are in strong contrast to (a) the growth of Urumqi's fleet of motor vehicles and (b) to the growing number of stable regimes of Urumqi's boundary layer climate during same period. It is concluded that the (regional and) local administrative technical countermeasures have efficiently lowered Urumqi's air pollution levels.     

Characterisation of Seed Lipids from Bixa orellana and Trachyspermum copticum

Bixa orellana L. seeds possess a resinous lipid (6.3 %), which has a pungent and spicy odour. The seed is known for its medicinal properties such as anti‐inflammatory, antipyretic activity and as a cure for tonsilitis. Trachyspermum copticum L. seed is a well known digestive aid and relief from colic pain. T. copticum possesses essential oil rich in thymol (>50 %) and lipid (15.6 %). The present study was aimed to quantify lipid classes of these two species by silicic acid chromatography and analyze their fatty acid composition by gas chromatography (GC) and gas chromatography–mass spectrometry (GC–MS). It was observed that the seed lipids are rich in neutral lipids with 98.1 and 95.2 % and lower quantities of glycolipids of 1.5 and 3.8 % and phospholipids of 0.36 and 1.0 % in B. orellana and T. copticum, respectively. The fatty acid composition of B. orellana seed lipid showed major quantities of palmitic (26.9 %), linoleic (26.1 %), oleic (17.5 %), linolenic (15.1 %), stearic acid (10.8 %) and small quantities of eicosanoic acid (3.6 %). In T. copticum seed lipids, petroselinic acid (68.3 %) and linoleic acid (25.3 %) together constituted 93 % of the total lipid. The results revealed that the lipids after recovery of the essential components namely, bixin and volatile oil from B. orellana and T. copticum, respectively can be further explored for industrial applications.     

基于UIO序列的类重要性度量

程序理解是测试和维护大规模面向对象程序的关键,选择程序的关键类优先开展分析是理解程序结构的一个有效的方法.为支持自动识别软件系统中的关键类,本文提出了一种基于UIO序列的类重要性度量方法.首先将软件系统抽象为一个以类为转换的有限自动机模型,随后求解该自动机的UIO序列,将该序列集合转化为状态转换树.通过递归计算状态转换树的节点复杂度求得类重要度.并在考虑异常传播的基础上改进了算法.最后通过实验验证了算法的有效性.     

基于多特定决策类的不完备决策系统正域约简

现有的属性约简方法大部分关注决策系统中的所有决策类，而在实际决策过程中决策者往往仅关注决策系统中的一种或几种决策类。针对上述问题，提出基于多特定决策类的不完备决策系统正域约简的理论框架。首先，给出不完备决策系统单特定决策类正域约简的概念；第二，将单特定决策类正域约简推广到多特定决策类，构造了相应的差别矩阵及区分函数；第三，分析并证明了相关定理，提出基于差别矩阵的不完备决策系统多特定决策类正域约简算法（PRMDM）；最后，选取4组UCI数据集进行实验。在数据集Teaching-assistant-evaluation、House、Connectionist-bench和Cardiotocography上，基于差别矩阵的不完备决策系正域约简算法（PRDM）的平均约简长度分别为4.00、13.00、9.00和20.00，PRMDM算法（多特定决策类中决策类数目为2）的平均约简长度分别为3.00、8.00、8.00和18.00。实验结果验证了PRMDM算法的有效性。     

Analytic inversion for a general case of certain classes of matrices

The explicit inverse and determinant of a class of matrices are presented. The class under consideration is defined by 4n ? 2 parameters, analytic expressions of which form the elements of the upper Hessenberg type inverse. These analytic expressions enable recursion formulae to be obtained, which reduce the arithmetic operations to O(n ²). The Hadamard product of two specific structures' matrices provides the class presented, special cases of which are already known classes of test matrices.     

昭通晚清民国时期彝族上层社会建筑概貌及分析

昭通是彝族的重要发祥地之一，彝族自古以来对昭通的政治、经济、文化影响极大，彝族上层社会的建筑为研究昭通彝族文化和历史的变迁提供了重要的实物证据。文章通过昭通晚清民国时期彝族上层社会建筑的调查，对典型建筑案例进行了概述，分析了昭通晚清民国时期彝族上层社会的建筑特征和建筑文化，希望地方历史建筑的保护能够得到加强和重视。     

PH的一种复杂性类分解

本文基于ΔPK-复杂性类给出多项式时间谱系PH的一个分解,并讨论了相关的一些性质。利用该分解给出PH是否只有有限个层次这一重要计算复杂性理论问题的两个充分条件,并证明了NP中稀疏集构成的语言类在LP2∧中。