查询词典 data point

In this framework, image frame, audio and text are represented, which are the three modalities in video shots as data points by the 3rd-order tensor. Then a subspace embedding and dimension reduction method is proposed, which explicitly considers the manifold structure of the tensor space from temporal-sequenced associated co-occurring multimodal media data in video. It is called TensorShot approach. Transductive learning uses a large amount of unlabeled data together with the labeled data to build better classifiers. A transductive support tensor machines algorithm is proposed to train effective classifier. This algorithm preserves the intrinsic structure of the submanifold where tensorshots are sampled, and is also able to map out-of-sample data points directly. Moreover, the utilization of unlabeled data improves classification ability.

在此框架中，视频镜头首先被表示成由视频中所包含的文本、视觉和听觉等多模态数据构成的三阶张量；其次，基于此三阶张量表达及视频的时序关联共生特性设计了一种子空间嵌入降维方法，称为张量镜头；由于直推式学习从已知样本出发能对特定的未知样本进行学习和识别，最后在这个框架中提出了一种基于张量镜头的直推式支持张量机算法，它不仅保持了张量镜头所在的流形空间的本征结构，而且能够将训练集合外数据直接映射到流形子空间，同时充分利用未标记样本改善分类器的学习性能。

Due to the development of computer and the popularization of Internet, more and more enterprises have adopted the computer network for the data communication to achieve the commercial business transaction or factory control. However, the data transmission in network is complicated and unforeknown. Thus, it will get some problems or affect the system operation if data are lost during the data transmission. The TCP/IP protocol is currently one of the most popular communication protocols. And the service quality from user request will be higher and higher. Therefore, How to provide the high–quality data communication and avoid the data loss is the goal of this thesis.

随著电脑技术的进步以及电脑网路的普及，愈来愈多企业体系利用电脑网路来做资讯交换，以达到商业交易或工厂控制的目的，而网路传输状况难以预测，若在资料传输过程中发生资料遗失，小则造成资料不对，大则造成应用系统无法正常运作，现今工商业对网路应用的依赖度愈来愈大，稍有闪失，都有可能招致莫大的损失， TCP/IP协定是目前网际网路上最被广泛使用的通讯协定，而目前使用者对其服务品质的要求也愈来愈严格，因此，如何在现有的网路下避免资料遗失及提高通讯品质，就是本篇论文所要探讨的课题。

The coding device for DVD system and other record systems includes audio and video coder to convert the input audio and video data into audio and video data stream; system coder to convert the audio and video data stream into audio and video data stream of, say, MPEG format; audio and video data stream buffer before the system coder to store audio and video data stream generated during one set period by the audio and video coder; and buffer downstream the system coder with relatively low capacity and capable of performing the record operation of pre-trigger.

提供了在DVD系统等记录系统中使用的编码设备。该编码设备包括分别把输入音频和视频数据转换成音频和视频数据流的音频和视频编码器、把音频和视频数据流转换成例如MPEG格式的音频／视频数据流的系统编码器。还包括系统编码器上游的音频和视频流缓冲器，它们分别存储音频和视频编码器在一段预选的时间内产生的音频和视频数据流，从而使系统编码器下游的缓冲器的容量较小并有效地进行预触发器记录操作。

This paper is based on data mining technology,use weka as a library data mining software tool,use weka's j48 tree algorithm and data association analysis library data.mining useful data which user needs from the mass library data and get reasonable results.ultimately to improve work efficiency and scientific management.

该文以数据挖掘技术为基础，利用weka软件作为图书馆数据挖掘工具，通过weka里的j48树算法和数据关联等算法，对图书馆的馆藏数据进行相应的分析，从海量数据中挖掘出用户需要的有用数据，并得到合理的统计结果。最终达到提高工作效率，能够科学管理的目的。

We can divide web data into two kinds: structural data and unstructured data. We have maturer methords to deal with structural data. However, because traditional database bottom can not deal with unstructured data, a wey that deal with unstructured data need be presented.

对于结构化的Web数据，已经有较为成熟的解决方法；而对于非结构化的Web数据，由于传统数据库的底层问题，不能用来处理非结构化数据，迫切希望能提出一种方法进行非结构化数据的处理。

Firstly, in this paper the importance and necessity for the land using data sharing have been expatiated amply. The land using data sharing, which refers to two main concepts, has been summarized. The two concepts are acquisition and processing of land using data as well as application of land using data. In WebGIS, there are two corresponding roles, namely, data yielder and data user or consumer.

本论文首先详细阐述了土地利用数据共享的重要性和必要性，并总结土地利用数据共享活动涉及了两个主要概念：土地利用数据的获取与处理、土地利用数据的应用，在WebGIS中即对应两个角色：数据提供者和数据使用者即用户。

The design includes many new technologies in data exchange and data security, they are: Converting non-XML data to XML in order to apply the XML security technology to more data types; Using the technology of XKMS and PKI to manage the public key in order to overcome the shortage of complexity and difficulty in using PKI and improve the efficiency of the public key management; Using Diffe-Hellman algorithm combining with the public key technology to exchange secret key in order to furthest increase the security of the key; Using the technology of XML digital signature algorithm to ensure the integrity, message authentication and signer authentication of the data; Using the technology of XML encryption to ensure confidentiality of the data; Defining delegable access rules on the structure and content of the XML documents to provide fine-grained access control. According to the XML security framework, a XML security library is designed and implemented.

在对XML及其安全技术主要包括XML加密技术、XML数字签名、公钥管理、密钥产生和交换进行研究的基础上，结合传统密码技术和XML安全技术提出了一种基于XML的数据安全模型，模型中包含了目前网络数据交换和安全领域的一系列最新技术，主要有：将非XML格式数据转换为XML格式数据的方法，以保证现有的大部分网络数据能够应用XML安全技术；使用XML公钥管理技术结合PKI技术管理公钥，以克服PKI技术实施和部署复杂、难度大等问题，提高公钥管理的效率；使用Diffie-Hellman算法结合公钥技术交换密钥，以最大限度保证密钥的安全；使用XML数字签名技术保证数据的完整性、不可否认性和发送方身份认证；使用XML加密技术保证数据的机密性；根据XML文档的内容和结构定义授权访问规则进行良好的访问控制。

Under the guidance of DPRIF integrating strategy, we set up an integreted SOFM-SVM model. Then we analyzed and optimized the model from the aspects of operating mechanism, data interface and function expandedness. The PCA method was introduced to reduce dimensionality and extract features, then to strengthen the clustering explanation; defined a CMI index, to ascertain the most effective or the best clustering number; A new Anti-NO algorithm was proposed to recognize and to filter the suspecious data in the sample; The medium result of SVM model was used to extract the borderline datas between two classified groups. This research compensated for the achievements of data recognition including pattern data, noises data and borderline data.

在DPRIF整合策略指导下构建了一个整合的SOFM-SVM模型，对该模型从运作机制、数据接口、功能扩展几个方面进行分析和改进：引入PCA方法进行数据降维和特征提取，用以加强聚类解释；结合统计聚类中的聚类误差概念定义一个聚类数有效性指标，利用SOFM算法中间结果进行指标求解，以筛选出有效或最佳聚类数；提出一种新的噪声识别算法用于对样本中的异常数据进行甄别和过滤；利用SVM模型的中间结果提取分类边界数据；进一步充实了包括模式类、噪声集和边界在内的模式识别成果。

Facing the rapid incensement of large data collection, enterprises should have potent data analysis tools to transform the "fluent data" into "valuable knowledge", otherwise, large data will become "fluent data without useful information", what we call "data grave".

面对快速增长的海量数据收集，企业必须要有有力的数据分析工具将&丰富的数据&转换成&有价值的知识&，否则大量的数据将成为&数据丰富，但信息贫乏&的&数据坟墓&。

Furthermore, it can not define indiscernibility relationship when default data is included in rough data. In this dissertation, a new rough set model, which can be used to process original data which include default data, fuzzy data, or incomplete data, is proposed based on clustering methods.

本文从实际应用中存在缺值、不完全和噪音数据等问题为出发点，结合人机交互技术中数据挖掘的应用，研究分析了数据挖掘中的一些理论与方法，取得了以下主要研究成果： 1。

In contrast to the ubiquitous rising-sun-with-rays military flag of the Japanese, Chinese banners and ensigns feature a range of designs.

与遍地都是的太阳军旗不同，中国人的旗帜和徽章设计得各式各样。

From their small corner of Feng's Guangzhou headquarters -- a jumble of pink leashes, squeezable rubber steaks, and plastic doggy Santas for Fido's stocking -- Soleil's designers come up with at least five new products a month.

从Feng 设在广州总部的产品展示柜台上可以看到，Soleil的设计师每月至少设计出5件新产品。