融合La格虚词语义信息的藏文La格分类模型

采用深度学习方法实现藏文La格(■)分类是一项具有挑战性和重要研究意义的藏语自然语言处理任务.藏文La格的自动分类更加依赖于上下文语义信息和特征的时序性,该文通过分析La格虚词的用法及语义特征,在设计La格虚词语义信息标记算法的基础上,提出一种融合La格虚词语义信息的藏文La格分类模型.该模型首先以每个音节及对应La格虚词或其它音节的语义特征嵌入作为输入,丰富嵌入向量的语义信息,增加输入特征的多样性;然后采用一维卷积融合并学习每个音节及对应La格虚词或其它音节语义信息的局部特征向量,提高卷积层的空间特征学习能力;其次使用双向长短时记忆网络(long short-term memory)(LSTM)学习时序特征,提高时序特征的学习能力;最后使用注意力机制对双向LSTM层每一时刻的输出特征进行加权融合,充分利用每一时刻的输出特征,以提高最终文本表示的特征质量.在TLD藏文La格数据集上的实验结果显示,该模型的分类效果优于基线模型及仅用藏文音节嵌入的模型,在测试集上的分类准确率为93.10%.

现代汉语格的层级系统及其分类描写

从语法研究的三个平面来看,格属于语义平面。它存在于句模结构中,指名词性成分跟动词组成句模结构时所担当的语义角色,格关系不因表层句法结构的不同而发生改变。格是一个有层级的系统,格的类别主要由不同类别的动词规定。

现代汉语格系统的层级分类

从语法研究的三个平面来看 ,格属于语义平面。它存在于句模结构中 ,指名词性成分跟动词组成句模结构时所担当的语义角色 ,格关系不因表层句法结构的不同而发生改变。格是一个有层级的系统 。

基于属性分类的悲观概念格与乐观概念格

结合实际需求,在给定属性分类的形式背景中,首先定义悲观分类形式背景和乐观分类形式背景及其算子与概念,研究它们与原形式背景的算子、概念之间的关系.然后,对于悲观分类形式背景,建立原概念格与悲观分类概念格之间的映射,给出由原概念格直接生成悲观分类概念格的方法.对于乐观分类形式背景,引入概念包含映射,研究原概念格与乐观分类概念格之间的关系,给出对应的概念格生成方法.最后,通过例子阐述悲观分类概念格与乐观分类概念格在实际问题上的应用及语义解释.

现代维吾尔语格的分类

该文认为:维吾尔语共有10个格,并按主要语法功能将它们分为五类;通过对定格词和定状格词替代功能的分析,诠释了维吾尔语一个词中多种词尾叠加现象的成因。

The Classification of Positive Definite Unimodular Lattices Over Z[(1+21^(1/2))/2]

In this paper, the author applies adjacent lattice method and Siegel mass formula to determine the classes of positive definite unimodular lattices of rank 4 over Z , and obtains that the class number of unit genus gen( I 4 ) is nine and the class number of even unimodular lattices is three, and also gives the representative lattices of each class.

超级模数博弈的存在性

定义了一类在有序Banach空间上的超模博弈,并利用著名的Birkhoff不动点定理证明了有序Banach空间上超模博弈Nash均衡的存在性.

Structure identification for compound I separated and purified from taxoids-produced endophytic fungi (Alternaria. alternata var. taxi 1011)

Endophytic fungi are widely found in almost all kinds of plants. Many endophytic fungi can produce some physio-logical active compounds, which are same to or analog to those isolated from their hosts. Producing physiological active com-pounds through microbial fermentation can give a new way to resolve resource limitation and to find out alternative source. Through the methods of organic solvent extraction, thin layer chromatography (TLC) and column chromatography, compound I was isolated, purified from the liquid fermentation metabolites of the taxoids-produced endophytic fungi (Alternaria. alternata var. taxi 1011 Y. Xiang et LU An-guo) that was screened from the bark of Taxus. cuspidata Sieb.et Zucc.. Compound I was identified as one kind of taxoids type III, based on the analyzing results by using the methods of ultraviolet spectroscopy (UV), infrared spectroscopy (IR), mass spectrometry (MS) and nuclear magnetic resonance spectroscopy (NMR). This study provides a com-pleted method for separation and purification of the endophytic fungi as well as structure identification of its fermentation me-tabolite

Distribution patterns of benthic diatoms during summer in the Niyang River,Tibet,China

The aim of this study was to investigate the distribution, density, community structure and biodiversity characteristics of benthic diatoms, and to analyze whether differences in species composition and abundance exist in different regions of the Niyang River, Tibet. Among the 157 taxa observed in 15 sampling sites in the main river and tributary, most were casual species （〉100）, the relative abundance of the genera Achnantkes and Fragilaria was 67% of the total relative abundance. Acknantkes minutissima was the most important species and dominated the whole river reaches （average relative abundance was 30%）; the average diatom densities were 7.4 x 105 cell/cm2 at all sites, and increased slowly from the upper section to downriver. The significant indicator taxa with higher relative abundance were Achnanthes biasolettiana （18.0%） and Fragilaria arcus （18.2%）, Fragilaria capucina var. vaucheriae （31.2%）, Fragilaria construens var. renter （11.3%） and Cymbella affinis （11.0%） in the upper, tributary and mid-river sections, respectively. Achnanthes minutissima was the most abundant species （56%） in the downriver section. Biodiversity indices showed a gradual decrease from the up- to down-river section, and dominant species were more abundant in the upper and mid-river sections than in the downriver section. A two-way indictor species analysis （TWINSPAN） of diatom composition clearly showed four different groups, namely the upper, mid, lower and tributary sections. Detrended correspondence analysis （DCA） supported the results of TW1NSPAN, and the characteristics of site distribution and species composition in the Niyang River supported the spatial structure of diatom assemblages. This study indicates that bio-assessment programs utilizing benthic diatoms could clearly benefit lotic water with regional stratification.

TWIN SUPPORT TENSOR MACHINES FOR MCS DETECTION

Tensor representation is useful to reduce the overfitting problem in vector-based learning algorithm in pattern recognition.This is mainly because the structure information of objects in pattern analysis is a reasonable constraint to reduce the number of unknown parameters used to model a classifier.In this paper, we generalize the vector-based learning algorithm TWin Support Vector Machine(TWSVM) to the tensor-based method TWin Support Tensor Machines(TWSTM), which accepts general tensors as input.To examine the effectiveness of TWSTM, we implement the TWSTM method for Microcalcification Clusters(MCs) detection.In the tensor subspace domain, the MCs detection procedure is formulated as a supervised learning and classification problem, and TWSTM is used as a classifier to make decision for the presence of MCs or not.A large number of experiments were carried out to evaluate and compare the performance of the proposed MCs detection algorithm.By comparison with TWSVM, the tensor version reduces the overfitting problem.

Taxonomic Studies on a New Marine Ciliate, Apocoleps magnus gen. nov., spec. nov. (Ciliophora, Colepidae), Isolated from Qingdao, China

The morphology and infraciliature of a new marine colepid ciliate, A＇pocoleps magnus gen. nov., spec. nov., are de- scribed based on living observations and silver impregnations. The new genus Apocoleps is characterized by having 8 （vs. 6 in most other related genera） armour tiers, spines at both ends of the cell, 3 adoral organelles and plates with 4 reniform uni-windows. Apo- coleps magnus spec. nov. is defined by the following features： body elongated and slightly curved, about 100-120p.mx 3545 lam in vivo; anterior tertiary tier plate with four uni-windows, most secondary and main tier plates with four uni-windows, posterior tertiary tier plate with two uni-windows; left plate margin slightly serrated; on average 23 transverse and 22 longitudinal ciliary rows; one terminal contractile vacuole; marine habitat.

基于曲率值追踪的气象槽线自动分析算法

针对传统槽线自动分析算法中存在的处理效率较低,且分析准确度不高的问题,提出一种基于曲率值追踪的槽线自动分析算法。算法利用位势高度场格点数据,直接计算格点处曲率值,并根据曲率值及相对位置对格点进行分类判别,以提取候选槽点;在候选槽点中筛选合理槽点进行追踪连接;最终通过后处理步骤,得到槽线自动分析结果。实验结果表明,该算法可有效实现气象槽线的自动分析,在分析准确度和运行速度方面较之前算法均有所改进。

Regional Economic Types and Development Strategy of Small Towns in Suburb Beijing

The purpose of this article is to classify the regional economic types and development dynamics of small towns in suburb Beijing by applying the classification methods developed by Nelson's method and Location Quotient method into 183 small towns in Beijing.Four types of small towns are thus identified,including urban agriculture dominated towns,manufacturing dominated towns,service industry dominated towns and comprehensive type towns with balanced economic development.Within the environment of geographical information system,the spatial distribution pattern of four types of small towns with their evolution trend is analyzed.The results indicate that four types of small towns have obvious'core-periphery'spatial structure but with different functional orientations and evolution mechanisms.Based on this,the different development strategies for each type of small towns are summarized,providing a scientific reference for the differentiating planning and development strategies of these small towns.

Research on the credit classification of practicing qualification personnel in construction market based on self-organizing neural network

Combining with the characters of the practicing qualification personnel in construction market,evaluation method based on the self-organizing neural network is brought out to analyze the credit classification of the practicing qualification personnel. And the impact factors on the credit classification of the practicing qualification personnel,such as the number of neurons,the training steps,the dimension of neurons and the field of winning neurons are studied. Then a self-organizing competitive neural network is built. At last,a case study is conducted by taking practicing qualification personnel as an example. The research result reveals that the method can efficiently evaluate the credit of the practicing qualification personnel;thus,it could provide scientific advice to the construction enterprise to prevent relevant discreditable behaviors of some practicing qualification personnel.

Morphological and Structural Categories： Geometric Attributes of the Form of Places

The author knows how to interpret the city is a complex task that met the physical form and also all the images that are transmitted by the sensations. However the aim of this paper is to demonstrate the importance of studying the morphological categories, using the geometric attributes of the shape of places, facilitating the development of three-dimensional view of the architects, in particular for students of architecture. This article is part of the Project ＂Education of the Eye＂ which introduces the training of the eye as a strategy to understand the architectural forms characterizing the volumetric space （architectural, urban and landscaping）, through the polyhedra and their relationships. It is linked to the research line Teaching of Architecture Graduate Program in Architecture -- PROARQ-FAU/UFRJ. The investigation of the way in this direction, interacts with the knowledge produced by the groups SEL-RJ -- Research Group on Systems of Open Spaces in Rio de Janeiro, ProLUGAR -- Design and Quality of Place and EAG-Education-Environment Group.

Heterogeneous Responses of Chinese Cities' Housing Prices to Monetary Policies

This works examine the responses of housing prices to the monetary policies in various Chinese cities. Thirty-five large and medium sized Chinese cities are classified into six clusters applying the minimum variance clustering method according to the calculated correlation coefficients between the housing price indices of every two cities.Time difference correlation analysis is then employed to quantify the relations between the housing price indices of the six clusters and the monetary policies.It is suggested that the housing prices of various cities evolved at different paces and their responses to the monetary policies are heterogeneous,and local economic features are more important than geographic distances in determining the housing price trends.

Classification and Regression Methods with Data Mining Algorithms

This article deals with implementation of the classification and regression trees into the DMAIC phases of Six Sigma methodology. Six Sigma methodology seeks to improve the quality of manufacturing process by identifying and minimizing variability of this process. Using the classification, regression and segmentation trees as a part of the Data Mining methods could improve results of DMAIC phases. This improvement has a direct impact on the Sigma performance level of processes. The author introduces research results of implementation Data Mining algorithms into retail sales promotion. The author implements classification and regression techniques in our research. As a software tool has been selected SPSS PASW Modeler. The author deals with more data mining algorithms ad their implementation in the DMAIC phases. The article is divided into several parts. The first part is the introduction to Six Sigma methodology, the second deals with classification and regression trees. The third part describes tree research focused on the implementation of data mining algorithms and the fourth section summarizes the research results.

共享经济商业模式的分类和理论模型——基于三个典型案例的研究

为了加深对共享经济商业模式的认识,本文首先回顾了共享经济的相关研究,提出共享经济的概念应包括主体、客体、途径、目标4个构成要素,并从提高资源利用效率和提高社会福利的角度界定了共享经济的概念。其次,按照表现形式和主体类型两个维度,提出了"九宫格共享经济分类模型",将共享经济分为9种基本类型,并应用该分类模型对中国当前的共享经济业态进行了分类。再次,构建了"PPC三维网络共享经济商业模式理论模型",应用该模型分析了摩拜、知乎、三一重工设备共享平台的商业模式,并提出了相应的建议。最后,通过案例研究证明了该理论模型的科学性和普适性,并对共享经济未来的发展趋势进行了展望。

Static Model Classification Status： Taking Into Account Emerging External Factors

Analysis of the problem of predicting bankruptcy shows that foreign and domestic models included only internal factors of enterprises. But the same indicators of internal factors in the rapidly changing external environment can lead to bankruptcy, and not in others. External factors are the most dangerous, because the possible influence on them is minimal and the impact of their implementation can be devastating. This paper focuses on the same factors to assess the impact of the macroeconomic indicators （extemal factors） on the parameters of static models predicting a local approximation of the crisis at the plant. To accomplish the purpose, a Spark set of 100 companies was compiled, including 50 companies which officially declared bankruptcy in the period of 2000-2009 and 50 stable operating companies with a random sample of the same time period. External factors were extracted from the Joint Economic and Social Data Archive1 The author compared two data sets： （1） microeconomic indicators--money to the total liabilities, retained earnings to total assets, net profit to revenue, Earnings Before Interest and Taxes （EBIT） to assets, net income to equity, net profit to total liabilities, current liabilities to total assets, the totality of short-term and long-term loans to total assets, current assets to current liabilities, assets to revenue, equity to total assets, and current assets to revenue; and （2） external factors--index of real gross domestic product （GDP）, industrial production index, the index of real cash incomes, an index of real investments, consumer price index, the refinancing rate, unemployment rate, the price of electricity, gas prices, oil price, gas price, dollar to ruble, ruble euro Standard ＆ Poor （S＆P） index, the Russian Trading System （RTS） index, and region. The aim of the comparison results paging classes ＂insolvent＂ and ＂non-bankrupt＂ is achieved using two methods： classification and discrimination. In both methods, computational procedures are realized with the use of algorithms linear regression, artificial neural network, and genetic algorithm. In the 2-m model, data set includes both internal and external factors. The results showed that the inclusion of only the microeconomic indicators, excluding external factors, impedes models about two times.

Remote sensing-based artificial surface cover classification in Asia and spatial pattern analysis

Artificial surfaces, characterized with intensive land-use changes and complex landscape structures, are important indicators of human impacts on terrestrial ecosystems. Without high-resolution land-cover data at continental scale, it is hard to evaluate the impacts of urbanization on regional climate, ecosystem processes and global environment. This study constructed a hierarchical classification system for artificial surfaces, promoted a remote sensing method to retrieve subpixel components of artificial surfaces from 30-m resolution satellite imageries(Globe Land30) and developed a series of data products of high-precision urban built-up areas including impervious surface and vegetation cover in Asia in 2010. Our assessment, based on multisource data and expert knowledge, showed that the overall accuracy of classification was 90.79%. The mean relative error for the impervious surface components of cities was 0.87. The local error of the extracted information was closely related to the heterogeneity of urban buildings and vegetation in different climate zones. According to our results, the urban built-up area was 18.18×104 km2, accounting for 0.59% of the total land surface areas in Asia; urban impervious surfaces were 11.65×104 km2, accounting for 64.09% of the total urban built-up area in Asia. Vegetation and bare soils accounted for 34.56% of the urban built-up areas. There were three gradients: a concentrated distribution, a scattered distribution and an indeterminate distribution from east to west in terms of spatial pattern of urban impervious surfaces. China, India and Japan ranked as the top three countries with the largest impervious surface areas, which respectively accounted for 32.77%, 16.10% and 11.93% of the urban impervious surface area of Asia. We found the proportions of impervious surface and vegetation cover within urban built-up areas were closely related to the economic development degree of the country and regional climate environment. Built-up areas in developed countries had relatively low impervious surface and high public green vegetation cover, with 50–60% urban impervious surfaces in Japan, South Korea and Singapore. In comparison, the proportion of urban impervious surfaces in developing countries is approaching or exceeding 80% in Asia. In general, the composition and spatial patterns of built-up areas reflected population aggregation and economic development level as well as their impacts on the health of the environment in the sub-watershed.