Search

This paper presents a method of utilizing background corpus and dictionary to calculate similarity between unknown words. In the method, the best concept expression of unknown word in corpus was obtained from the background of it, then constructed context for the best concept expression. The connotation meaning of unknown word was determined by the difference between the context of the best concept expression and its own context. The similarity between unknown words was calculated by utilizing semantic dictionary. This method avoids the problems of mistaken segmentation and abused segmentation, which exist in the traditional method of calculating similarity between unknown words, which is based on segmentation strategy. Experimental results show that the method proposed in this paper is high effective.

The cost-model for XML stream is the theory foundation in the research of XML stream dissemination, fragmentation and evaluation. According to the matching operations in fragmented XML stream system, this paper analyzes the features of the system overhead on client, network and server, and brings in the cost model for fragmented XML stream based on Hole-Filler model. Based on the cost model, this paper illustrates the influence of the fragmentation policy on XML dissemination system by an application and proposes a series of selection policies for XML stream fragmentation to improve the system efficiency, adaptivity and scalability.

Choose of training samples is an important topic for the support vector machines applied research. This paper proposes a new image segmentation method: the two-dimensional histogram of the segmented image is used to guide to select training samples to train the classical support vector machines. First of all, calculate the two-dimensional histogram of the segmented image and then get its two thresholds by the rough estimation which proposes in the paper; then, according to these two thresholds to find the range to choices the training samples; finally, choose the appropriate number of training samples in the certain range to train support vector machines. The method have better results whether in visual or in quantitative analysis.

With the rapid development of information technologies and the rapid spread of information systems, security of data storage has become an important issue in information systems. This paper introduces the data encryption technologies of encrypting file system (EFS) and traditional full-disk encryption (FDE), and points out the problems of data encryption of EFS and FDE. Combined with the features of trusted plateform module (TPM), this paper constructs a trusted full-disk encryption (TFDE) based on TPM. In the model, the hard disk is divided into four partitions, and each partition is encrypted with different keys, and the data of different users is encrypted with different keys too. The model can not only ensure that the data can't be stolen when the disk is lost or stolen, but also ensure that the data can't be leaked between different users. Finally the security of the model is demonstrates in depth.

By extending Cournot model, this paper explores the impact of open source software (OSS) to the price, output and profit of proprietary software (PS) and social welfare in a software market. In this paper, we assume that proprietary software producer aims at maximizing profit, the open source software is freely available and the market exhibits network externality. We find that the appearance of open source software in a software market monopolized by the proprietary software doesn't necessarily lower the price of proprietary software, decrease the output and profit of proprietary software producer and increase the level of social welfare, what happens in reality depends on the learning (maintenance or development) cost of open source software.

This paper analyzes the web logs by using data mining technology. Through structure mining, usage mining and content mining, analyzes and studies the existing problems in current mining technology. Aiming at relevant key preprocessing schedule such as Frame page filtration, time-out threshold value setting and long-time setting, etc. this paper puts forward a modified data preprocessing technical method and compares the mining results through tests before and after the modification. Experiment proves that, the modified preprocessing technology is feasible and it can solve problems existing in relevant preprocessing effectively.

The emergence of cloud computing was the result of the development of parallel technology, software technology, network technology. Cloud computing ultimate goal was to service users. Cloud computing could provide users with on-demand computing power, storage capacity and application service capability. As a new model of Internet applications, cloud computing could be the dominant mode to get services and information for us. At present there was no accepted definition of cloud computing, the paper described the definition of cloud computing from the user's point of view. And the paper generally introduces the current cloud computing platforms, the development, prospects and challenges of cloud computing.

With the rapid progress of urbanization, urban planning is attracted much more attentions. The popularization of computer technology and GIS technology provides the technical support for urban planning. This paper introduces the concepts of spatial analysis functions of GIS briefly, and analyses the spatial analysis function of the application in urban planning specifically. How to apply the GIS spatial analysis function in the different urban planning field is demonstrated in the paper, which also gives the direction of the future application of GIS science and technology in urban planning.

This paper introduces the applications and implementations of HGML technology in geo-browser side of G/S mode which is a new type of spatial information network access mode. First, on the basis of HGML technology research, pointed out its role in the geo-browser, and then summarized many advantages of using HGML as the core by analyzing the operating mechanism of the geo-browser, also pointed out that the support of HGML on the client aggregate services. At last, this paper cited the practical application of the geo-browser using HGML in many fields.

This paper is based on the “China Equipment and Education Resources System” project, aiming to share large equipment resources and related services nationwide. Based on the existing resource and service integration solutions, a new resource-service registry based architecture is proposed in this paper. The new architecture consists of two parts: resource registry and service registry. A service query optimization algorithm is proposed to improve the efficiency of the architecture. To validate the feasibility of the architecture, a prototype system is developed and the running effect is given.

This paper introduces the dissemination technical architecture design of large size satellite meteorological data for China Integrated Meteorological Information Service System(CIMISS) that assumed by National Meteorological Information Center(NMIC). In CIMISS project, it has two levels, one is national center, the other is provincial center. As the national level centre, NMIC assumed the responsibility of data collection and dissemination. Comparison to common material and data, Satellite files are usually much bigger especially when FY-3A satellite data be transmitted this year. If using existing communication system's architecture, the timeliness for provincial level to get satellite data could not be satisfied. In order to improve timeliness and reduce occupation of network bandwidth as much as possible, this paper designed a distributed ftp technical architecture which could combine with existing communication system well.

OINK is a C and C++ static analysis tool for Unix-like platform. It can be used for programmers to analyze and understand code, or can be used for software testers to test code and protect software quality. In addition, it is good at data flow analysis, and can be convenient to be re-used. This paper analyzes the implementation process of OINK in detail at first, and then studies the software migration skills in order to extend the application of OINK. Windows platform is widely used but quite different from Unix-like platform, and this paper also studies how to migrate the OINK from Unix-like platform to Windows platform.

Aimed at enhancing the adaptive information fusion capability of underwater target recognition in complex environment, we employed the decision fusion technology based on the Dempster-Shafer Theory. The information fusion model presented in this paper utilized neural networks and fuzzy theory, which is the basic theory for this model, to analyze the uncertainty of probe and apply it to information fusion of underwater target recognition. By fusing several target features, the model was able to improve the recognition ratio for targets. Through the treatment of experimental data, this paper validated that the method does increase the accuracy and validity of underwater target recognition and the credibility had been greatly improved.

Regarding the data collection process for a comprehensive automated environmental monitoring and precaution platform, this paper first analyses the key issues and challenges, including setting of monitoring targets, heterogeneity of monitoring equipments, diversity of communication methods, and online monitoring. Based on this, an embedded real-time environmental data collecting system is proposed. The contributions of the paper are as follows: we have designed an embedded environmental data collector which supports three collecting modes; we have proposed a novel integration technique for heterogeneous monitoring devices and an adaptive internetworking approach. The proposed system is a component of a distributed environmental monitoring platform which will be deployed in Guangzhou city.

In order to meet the sealing and strength requirements of the casing, tubing, and down-hole tools in Shengli oil field, Non-API thread are adopted, which include VAM-TOP thread and FOX thread. The types and advantages of the particular thread are introduced in the paper. The features and sealing design of the special threads are discussed. These two kinds of special threads are applied in about 50 wells of the Shengli oilfield, and the overall effects are good. However, the casing strings of 4 wells failed to seal in the recent operations, which caused serious damages to the productions of these wells. The finite element model of the VAM-TOP thread and FOX thread are developed by the CAE(Computer Assistants Engineering) technology in the paper to find the reasons of failure. The calculating results show the reason of the thread sealing failure is the thread button does not accord with the standards.

The paper puts forward the concept of collaborative mobile e-commerce by applying the theory of collaborative commerce to mobile e-commerce. For the security issues faced by collaborative mobile e-commerce, the paper proposes the wireless middleware based security model for collaborative mobile e-commerce. The model can first solve the security problems in the collaborative enterprise cluster, and meet requirements of mobile commerce as well. The model is well made to meet the security demand of the operation of collaborative mobile e-commerce.

This paper analyzes the development of the PageRank algorithm both domestic and international, expounds the practical application and development of the algorithm. On this basis, the dissertation analyzes the thinking of the traditional PageRank algorithm, sums up the advantages and disadvantages, and introduces several important PageRank improved algorithms and their thinking. The paper provides a time-feedback PageRank algorithm improvement for the shortcoming of the traditional PageRank algorithm focus on the old websites. Adding a feedback factor in the traditional PageRank algorithm can accelerate the old and useless Webs sinking, the new Webs rising, and taking into account the old and important webs not sink. Implementing the PageRank value is plus or minus with the time of webs. On the basis of theoretical research, implements the rank model based on the PageRank algorithm, the model contains the traditional web ranking system model and improved web ranking system model. Experiment on this model, contrasting the experimental results. Experimental results show that the improved algorithm can improve the satisfaction of the user's search results.

On the basis of analyzing characters of freeway network tolling data stream and SPEED algorithm, this paper proposes a data stream mining system using SPEED algorithm, and presents detailed design and implementation. The data stream mining system that presented in this paper includes three parts: data preparation, data stream mining and data Discard. The primary goal of our work is to establish a data stream mining system model designed for freeway network tolling at freeway's company. However this structure can be easily expanded to other areas of data stream mining.

This paper analyzed shortcomings of multi ingredients control methods. The four-group Ingredient system of certain marked cement was adopted to prove that, improper proportional control of each ingredient will lose quality. Aimed at solve the problem of limited samples, strong real time, unstable flow speed of raw meal, through reforming the mouth of input of raw meal, adding the control valve and using fuzzy support vector machine to predict and control the flow speed of raw meal, the result was better. Algorithm was proved and contrast experiments were applied with ANN and common kernel functions in this paper. Results give some kind of guide for similar questions.

To make the evaluation of search engine performance more scientific and systematic, this paper uses the Analytic Hierarchy Process method to build the search engine performance appraisal system. The paper presents the key indicators which impact the performance of Web search engine. By constructing discriminate matrix, computing the relative weighting and verifying the consistency we determine the index weights, and then calculates the search engine performance evaluation result. Finally, the article is validated through the application of the evaluation system, the result shows that the appraisal system is more scientific and rational, and it has a certain value.

The results of search engine usually contain a number of redundant information and how to eliminate it has become technology issues waiting to be explored. This paper proposed an improved elimination algorithm based on best similarity of the results. By analyzing the search results, extracting their key words, comparing the similarities, and clustering the results, the algorithm can eliminate the useless and reduplicate results. The experimental results show that the performance of the elimination algorithm in this paper has been much improved compared to spectral segmentation algorithm.

With advance in wireless and location technology, Location-Based Services (LBSs) become the main applications in pervasive environment. Privacy of user query is becoming an increasingly important issue. In our proposed technique, user's point location has been expanded to region location. Because LBSs can't distinguish the true user from the other users in same region, user query privacy is protected. This paper has analyzed the problems that exist in the protecting of query privacy in pervasive computing. On this basis, a double threshold privacy model based on user distribution matrix is proposed for protecting user query privacy, include query location privacy and query content privacy. In this paper, we present a user cloaking algorithm to dynamic generation and modification user cloaking region. Experimental results validate the efficiency and effective of proposed technique.

In this paper, the fragmentation of fuzzy numbers is defined as new sort indicators Based on the Centroid and fragmentation of fuzzy numbers, paper proposes a new ranking method of fuzzy numbers which can overcome the shortcomings of some ranking method to some degree. At last, paper illustrates the practicability of the proposed method by the analysis of examples.

This paper proposed a new visual data mining and knowledge discovery method based on Lorenz curve and choropleth map. Lorenz curve is used to discover numeric knowledge through graphic analysis objectively. The break values of classification system are determined adaptively through simplifying the Lorenz curve according to data distribution. Map readers can find and identify spatial distribution pattern on the choropleth maps under influenced by “Gestalt's Law” like continuity and closure imperceptibly. Two case studies are examined in detail in the paper. Data used here are population spatial data at the county level in 2000. Both of them verify the effectiveness and feasibility of the method in discovering spatial distribution knowledge of population and other social-economic variables.

This paper proposes a new reduction, that is, considers the relation of attribute reduction and attribute value reduction. When making reduction, both the attribute reduction and the attribute value reduction can be done simultaneously in the process of reduction in a decision table. In comparison with the traditional methods, the complicated comparison process in the reduction of the attribute value can be omitted. the number of the comparison times can be decreased and the reduction efficiency can be raised after using the new methods in this paper, at the same time, this paper gives some proofs.

The reliability of web services is a critical problem when use web service infrastructure, especially when the states of different web services are involved. This paper presents a new-designed reliable web service system through status monitoring and replication network. Traditional web service is regarded as stateless, though the state has great impact on the success rate of execution. In this paper, during the period of service execution, we maintain a list of functionally interchangeable service nodes to form a replication network. The states of the services in the replication network are well monitored through mechanism based on Web Services Resource Framework (WSRF). The component nodes of the replication network are carefully selected in consideration of both system reliability and the cost of maintain such a network. In this way, proper replica can be activated when current web service fail to fulfill user's demand and the success rate can be improved.

Data exchange is the kernel of distributed computing, and data sharing and exchange is one of the crucial problems in data systems over the wide area network. This paper proposes a new data exchange protocol based on Web Services, which defines service primitives and operation rules so as to provide a reliable transmission between both structured and unstructured data. Besides, it can resume transmission after break point. A universal data exchange platform based on Web Services is designed and implemented with respect to the proposed data exchange protocol, the transmission and adaptor mechanisms of which are also discussed. Furthermore, this paper proposes a new method using sliding windows for transmitting SOAP messages to improve the transmitting efficiency during data exchange, which would speed up SOAP message transmitting. Simulation results show that transmission efficiency and data exchange reliability are effectively improved.

By extending Hotelling game model, this paper explores compatible strategies between open source software (OSS) and proprietary software (PS). In this paper, we assume that PS producer aims at profit maximization, OSS producer pursues market share maximization and software user's expectations on network size may be fulfilled or myopic. For each type of user expectation, we study both cases of market fully covered and partly covered. The main findings are following: (i) in the case of myopic expectation, both OSS and PS producers choose maximum compatible degree in equilibrium. This conclusion is true in both fully covered and partly covered markets; (ii) in the case of fulfilled expectation, equilibrium compatible strategies depend on market coverage. When the market is partly covered, both OSS and PS producers choose maximum compatible degree in equilibrium. But when the market is fully covered, PS producer may choose maximum compatible degree or incompatibility as its equilibrium strategy (The choice depends on the learning (maintenance or development) cost of OSS, user's software development capacity and network externality); (iii) how the competitive strategies of OSS and PS producers are affected by the rival's compatible degree may be different in different user expectations (or market coverage).

As CIM subset, GIS data model of distribution grid was constructed by the relevant classes and relationships which were extracted from the CIM packages. The packages include Core, Topology, Wires, Meas, et al. Meanwhile, the spatial data exchange model of distribution network was discussed in this paper. Relational database technologies were applied to implement the spatial database, and generalization, aggregation and association relationships were converted to the relations of primary key and foreign key. The generalization mapping methods proposed by this paper can avoid the relationships of total model to fall on the leaf nodes, and then lower the relation's complexity among objects. Using object-oriented technology, the classes of GIS data model were extended by adding operations (Create, Update, Delete and GetProperties) and the necessary attributes. The above operations were utilized to generate commands which can maintain the data and relationships of spatial database. The Mount and Unmount operations were added to ConductingEquipment class, which can establish topological maintenance commands to maintain electrical relationships of conducting equipment. The approaches of class extending adopted by this paper make objects have responsibilities to sustain distribution network topology, and then cut down the maintenance difficulties of spatial database.

A design method of RBF neural network construction based on rough sets and orthogonal least squares (OLS) is proposed in this paper. First, the rough sets knowledge expression system is built taking use of the large number of system sample data. The impact of input variables on output variables is analyzed through computing an accuracy measure of input space knowledge on output space knowledge with reducing the input space of the knowledge expression system, System order and the input layer nodes of neural network can be decided by this way. Secondly, the hide layer nodes and the weights of output layer of the neural network can be obtained using OLS algorithm. Finally, the research of simulations on a nonlinear system are carried out using this method in this paper, the research results show the method is effective and feasible.

Storage of raster metadata is a key topic in spatial database. Although there are a few of abstract standards on raster metadata, there is not implement standard about it. This paper concludes three storage models implemented in current spatial databases and discusses their advantages and disadvantages. After that analyzing, the paper proposes a mixed storage method which is used the relational table to store structured metadata and used XML to store non-structured metadata, and gives its implementation solution.

The characteristics of distribution and burying of ground water are closely related to the nature of rocks and geological structures. The studies of this paper were based on spatial distribution law and structural features of the water control structures, combining with experiences of hydro-geological workers and expert knowledge. By using of the Geological Map and other auxiliary information, this paper presented the conceptual model and logic model of monoclinic water storage structure, as well as a set of automatic, semi-automatic identification mechanism of monoclinic storage structures. The results of this research can be referenced to modeling and identification of other types of water storage structures.

Currently, the e-government tends to take service as its core, while cloud computing has the characteristics of providing information technology in the way of service. This paper presents that e-government system can take advantages of cloud computing to change its function towards service, promote the industrial upgrading and push forward the green technology. We provide detailed analysis of the values and issues about the government, society and environment after the e-government system is moved to cloud computing environment. At the end of this paper there are the conclusions and expectation to e-government service platform based on cloud computing environment.

With the development of Scalable Vector Graphics (SVG) used in Location Based Services (LBS), the presentation and organization of spatial data based SVG became a burning question. The paper extracted the key technologies about presentation and organization of SVG_based spatial data according to the trend of development. The main points concluded acquisition of SVG data, presentation of attributes, construction of spatial data, and compression of SVG data etc. The paper also pointed out some current problems and research orientations in the future in some areas which concluded transform from multiple heterogeneous data to SVG data, presentation and construction of SVG_based data, and data compression etc. The paper can serve as a guide to used SVG in LBS.

Solar analysis function is the important indicator of modern civil construction and evaluation of environmental sustainable development, the increased buildings and the reduced space between buildings enhance the people's conflict of solar rights and residential environment problems. In order to safeguard the basic rights of the people and create a pleasant living condition for the citizen, it is necessary to put forward an effective method to measure the buildings' shadows, related solar analysis parameters, sunshine time, etc. The rapid development of 3D GIS, with its lifelike visual expression, powerful 3D analysis and operation functions, provides the solar analysis area a strong support. The traditional solar analysis software costs much on development, maintenance and modeling; has low independence and software portability and limited capability on function expansion. These limits couldn't match up to the need of true-3D GIS and the development of urban planning management. This paper aims at present the solar analysis module which based on plug-in technology and has successfully made it out. The thought of plug-in doesn't depend on specific software; the structure of “platform + plug-in” reduce the time of software development, giving the programmer more freedom to organize flexible functions. This paper has realized the accuracy and visualization requirement of solar analysis on the function: displaying the shadow of buildings, calculating sunshine hours and sunlight interval distance, drawing solar isochrones, solar conic surface and etc. At the end of the article, we show the results and which can clearly prove the advantage of this module.

In real-time and embedded systems, time is a very important feature, and time-triggered architecture is a strong candidate platform. However, because UML is built on event-triggered mechanisms, it is not suitable for time-triggered mechanisms. This paper introduces UML profile for MARTE, and presents some modeling mechanisms which describe the time-triggered features based on MARTE. Finally, the paper also elaborates on the application of these mechanisms through a real case.

This paper firstly discusses the problems of Web services application, proposes a method to address these problems to apply Virtual Web Services (VWS) which is an Integration of XML, Ontology, OWL-S, WordNet and other technologies, and then describes VWS mechanism, function and operation process. VWS can dynamically call real service and provide a unified service interface to the service requestor to meet the needs of users. VWS adopts double service matching engines to support more accurate and flexible service discovery. Finally, the paper describes eight core modules in detail to elaborate VWS system design and implementation process in E-commerce field.

This paper proposes a novel method for searching massive and small-scale XML documents via path constraints, referred to as XPC, to overcome drawbacks of conventional approaches. Firstly, we propose employing keywords with simple path constraints to retrieve XML data, which provides a user-friendly way without need of understanding complex knowledge and could express user demand accurately. This paper further proposes a novel method for computing term weight in documents via path constraints, called rtf-idf. It measures the similarity of path constraints by N-Gram and other factors according to the structure of the XML documents. Then we rank the relevant documents by an extension of the vector space model. The experimental results show that XPC indeed outperforms the baseline methods such as VSM in plain text and JuruXML.

Email has been one of the most important communication tools in modern society. Email spam has been around for more than a decade and its quantity greatly exceeds the ham. How to prevent spam spreading becomes an important problem and the anti-spam technology has become a hotspot both in research and industry area. This paper proposes a Re-Study Bayes method which uses information gain(IG) to calculate email characters and updates the characters set by re-study mode. By this method, the feature dimension is reduced, while the efficiency of extracting characteristics is highly improved. And the experiment in this paper gets more accurate results than previous methods.

In this paper we design and implement a Tibetan Topic Detection system to process the huge number of Tibetan language text on Web. It classifies the Tibetan text into several categories, performs clustering in each category to get the topic. According to the Tibetan grammar features, we give a Tibetan text clustering model TTCM (Tibetan Text Clustering Model) for the text from Internet news sites. We have studied the Feature representation, feature extraction, and clustering in the model separately. From the performed tests, It turns out that the text Clustering in this model has a good accuracy ratio and a good recall ratio. So it has high application value.

Now the traditional manual mode has already cannot satisfy modern management requirements for basic farmland and reserve cultivated land, Using computer technology, GIS technology can offer more convenient management tools and methods for land resources professionals. This paper, according to the principles of data standardization, technology modernization, scientific management, application of integrated technology, using the software design method, through in-depth analysis of basic farmland and reserve cultivated land management functional requirements, the basic farmland and reserve cultivated land management information system's overall architecture, database structure and functional structure were designed. It provided a platform design methods and ideas of inquiry, statistics and analysis system, in order to establish a platform with GIS and Web technology.

Clustering data streams is one of the important branches in mining data streams. Because of dynamic and massive characteristics of data streams, traditional data mining algorithms could not satisfy the requirement of online analysis and the appropriate value of number of clusters. The focus on data stream technologies is to design one-pass scan data set, and maintain an effective data structure in memory incrementally which is far smaller than the size of whole data set. In the paper proposes a new feature mining method named Gaussian Mixture Model with Genetic Algorithm (GMMGA), based on an extending method of Gaussian mixture model. This method is use a probability density based data stream clustering which requires only the newly arrived data, not the entire historical data. The GMMGA algorithm can determine the number of Gaussian clusters and the parameters of each Gaussian component through random split and merge operation of Genetic Algorithm. In the GMMGA, a function was made to threshold value to clusters to reduce the bad clusters effect on the clustering result. In this algorithm, it can improve the robustness and accuracy of the clustering numbers, also can save memory and run time. Experimental results show that the method is effective and has higher clustering precision compared with conventional STREAM algorithm and CluStream algorithm.

This paper will introduce why this engine is necessary, how to realize the engine, how to analyze the address data inputted, how to separate the input to every component, how to compare the input with truth data to judge if the input address can be delivered, how to standardize the output to make it professional and what the engine will return to the user to help users understand easily and clearly what happened when parsing the address.

A two-dimensional direction-of-arrival (2D-DoA) estimation approach based on radial basis function neural networks (RBFNN) is proposed in this paper. With the spatial cone angle, two RBFNN estimation models of two line-arrays of L-shape array are built respectively, which can estimate the spatial cone angle, namely dimension-degraded model. The intersecting line of two half-conical surfaces,which is corresponding to each spatial cone angle, is the arrival path of the unknown signal. Simulation results show that the proposed method can effectively reduce the training set, and the complexity of model building, it also has very high resolution, and effectiveness for future application.

Based on the analysis of mining prospecting and mining data, mine area three-dimensional model includes ground-based model, the exploration line profile model, the scope of mining model, mine body model, mining roadway model, reserve calculation block model and goaf model. This paper presents a kind of mine area modeling technique based on scanned raster images. This technology can collect data and establish three-dimensional model of the mining area rapidly based on “old data”, to realize three-dimensional visualization of the mining area and the dynamic management.

The paper presents a method for classifying Light Detection And Ranging (LiDAR) full waveform data, using an artificial neural network (ANN) approach. The ANN classifier used was a multilayer perceptron trained through the generalized predefined learning rule functions. Compared with the unsupervised classification based on Landsat 7 ETM+ (Enhanced Thematic Mapper Plus) images, the ANN classifier was suitable to better represent the nonlinearity in the LiDAR waveforms dataset. The multilayer perceptron neural network has proved to be a very effective tool for the classification of waveforms data. The classification results show that forest classification accuracy for broadleaved forest and needleleaved waveforms using ANN classifer is better than the classification accuracy of ETM+ image based-unsupervised classifer. Whereas, the overall classification accuracy of testing datasets using ANN classifier with using waveform data without a prior class probabilities is lower than the unsupervised classifier based-image.

With the rapid increase of web sources, more and more deep web databases become available. The information in these databases can only be accessed by submitting queries to back-end databases. However, the traditional search engine interfaces resemble extremely deep web interfaces. Therefore, it is difficult to distinguish them and to find deep web interfaces. This paper proposes a novel method of discovering deep web interfaces. We introduce a page division method to divide pages into separate parts. After that we remove the parts which don't contain search interfaces. At last we construct topic-specific queries to obtain results and distinguish deep web interfaces by analyzing the results. Experiment result shows that this method is effective and stable.

Along with engineering projects scale's expansion and projects quantity's increase, the owner more and more use EPC (Engineering Procurement and Construction) contract management pattern to raise the efficiency, reduce costs and transfer the majority of the project risks to the contractors. EPC contractors face bigger risks whose factors are complex throughout the processes of design, procurement and construction. Therefore, risk controlling has become an important part of the EPC contractor management. If the risk rank can be identified as soon as it appeared, followed by appropriate measures taken by the contractor, the loss of the contractor can be substantially reduced. In this paper, we combine the methodology of principal component analysis and neural network to choose projects statistical data that the contractor has completed as training data of the neural network. Then the EPC contractor risk early-warning system could be established. This system could divide the rank of the risk among the project process into three grades: slight risk, medium risk, and heavy risk, so that the contractor could make accordingly decisions in time based on the result of this system. Abstract—This electronic document is a “live” template.

Virtual Screening (VS), a productive and cost-effective technology in quest for novel lead compounds, is a new approach in the pharmaceutical industry. This kind of applications belongs to the family of large scale loosely coupled parallel applications, whereas concurrent tasks are independent of each other, and are most suited to run on the Grid environments. However, there are some important practical issues should be investigated, namely, the complicated scale of VS applications with diverse data sets, complexity of Grid Job Description Language and dynamic and heterogeneous grid environments. In this paper we propose a Comprehensive Task Management in the context of performing VS applications on Grid environments. First, we propose an extension of the Job Submission Description Language (JSDL) to specify the large scale data set packet strategy and a visual Task Editor to make the complexity of underlying task description transparent. Then we provide a coordinating mechanism to integrate multiple Grid platforms for VS applications. Furthermore, we present a task-level predictive model on estimating of execution time from historical information. Finally, we conduct a case study on discovering leads drug against Avian Influenza Virus H5N1 to highlight technical feasibility of our proposal.

In a sense, the agricultural product quality traceability system is regarded as one of the most important aspects in the construction of agricultural product quality supervision. By defining and delineating the relevant concepts, this paper aims to encourage some new approaches to accomplishing the connection of isolated enterprise information islands and the abstraction of agricultural product unit flow process at a higher level, and through exclusive coding, data collecting and data converging to establish a mode of agricultural product quality fast traceability system in China. At the same time, the requirement analysis on a fast traceability system leads to the study on the system structural design of the mode.