Search

This paper proposes a new classification model for management of resources in an educational repository. To facilitate usability of educational resources, a proper classification model should be provided. If no classification model, it causes several issues such as inefficient management of educational resources, incorrect retrieval, and low usability. To achieve the goal of this paper, the model should support a classification mechanism in multiple viewpoints. The proposed model enables more precise and systematic management of implementation resources, and also increases the ease of usability. [ABSTRACT FROM AUTHOR]

In this paper, we suggested a healthcare information management system for supporting healthcare home services. Our researching focus in this paper is how to design system and how to use constructed database on the framework for supporting healthcare home services. The healthcare database constructed by using this information for the purpose of healthcare home services is divided into the base information with real schemes and the context based information with view schemes. Finally, for verifying the practical use of healthcare information constructed in this paper, we show an example of healthcare home monitoring service using information, emergency call, home appliance control, and we describes the results of the experimental evaluation between the base and context based information in terms of execution service time for healthcare home applications. [ABSTRACT FROM AUTHOR]

In order to give adaptive instruction to the learner, it needs to know what knowledge the learner has and what goals the learner is trying to archive. This paper proposes a method to build a model based on Bayesian networks in order to assess the learner's knowledge level. An overlay student model based on Bayesian networks is presented. This student model built on knowledge relationships with prediction ability is discussed in details. In this model, an assessing method based on Logistic model in IRT with three parameters is provided to evaluate student's performance. A case study in the course of Data Structure is illustrated in this paper. [ABSTRACT FROM AUTHOR]

This paper introduces several feature selection methods on the VIP E-Mail account analysis, and then uses SVM to classify the users in order to maintain the ones who may leave in a short time. The websites hosting companies can follow the strategies, which are investigated in this paper, to reduce their labor cost and provide an attractive and simple registration for the users. The result shows that we can also reach the feasible predict accuracy. [ABSTRACT FROM AUTHOR]

Classification is an important task in the fields of data mining and pattern recognition. Now there have been many algorithms for this task, while most of them do not focus on the application in databases. In this paper we extend the definition of function dependency, prove the properties of the extended function dependency, and on this basis propose an algorithm for classification. According to the two theorems in the paper, our algorithm is complete that means it can find all the classification rules from the database. At last, we demonstrate our algorithm by an example that shows the validity of our algorithm. [ABSTRACT FROM AUTHOR]

OWL Web Ontology Language is a W3C recommended Language for representing machine-readable knowledge with more facilities than other knowledge representation languages. In this paper, we present a Use-Case based Fuzzy Ontology Constructing (UFOC) methodology for the building of edutainment ontology, which is encoded with OWL. Besides, a solution of representing fuzzy relation in OWL is also provided. In the edutainment ontology, two fuzzy relations, DE and AE are discovered, both playing vital roles in property calculation among online computer game. [ABSTRACT FROM AUTHOR]

In wireless sensor networks, an efficient Medium Access Control (MAC) protocol is critical, especially in terms of energy consumption and message latency. This paper introduces a novel protocol that nodes periodically sleep to conserve energy. When the source node has data packets to send, it uses a wakeup message to awake the sleeping nodes. The protocol needn't take the synchronization which must be achieved in S-MAC. It especially suits to the wireless sensor networks with narrow bandwidth. [ABSTRACT FROM AUTHOR]

The pest catastrophe prediction in ecology catastrophe is one of important part in expert system of intelligence agriculture and also the guarantee for preventing and controlling the pest catastrophe occurrence efficiently. The paper introduces the basic principles and methods of current catastrophe prediction in ecology catastrophe. According to modeling for the dynamic process of population, up-grown and amount of spawn, the paper finds out the trigger point and critical value inducing the pest catastrophe and implements the catastrophe prediction of Cnaphalocrosis Medinalis, which take measures before catastrophe to prevent big population coming into being. [ABSTRACT FROM AUTHOR]

In this paper, an integrated optimization model is developed to improve the efficiency of coal shipping. The objective is (1) to determine the types of ships and the number of each type, (2) to optimize the ship routing, therefore, to minimize the total coal shipping cost. Meanwhile, an algorithm based on two-phase tabu search is designed to solve the model. Numerical tests show that the proposed method can decrease the unit shipping cost and the average ship delay, and improve the reliability of the coal shipping system. [ABSTRACT FROM AUTHOR]

In this paper, we give a new definition of risk: a scene in the future associated with some adverse incident. The simplest scene is that, a stone is hooked over our head. When a wind comes, the stone would fall down on our head. In many cases, what can we do is to approximately represent risks. In this paper, we suggest an approach to infer risks with information described in natural language. With fuzzy logic, we give a sample to verify that the suggested approach is more flexible and effective. [ABSTRACT FROM AUTHOR]

With the longstanding development of Chinese petroleum enterprises, the control of discovery and exploitation goal cost is becoming a essential problem. Data Envelopment Analysis (DEA) is a linear programming method for measuring the relative efficiency of decision making units where the presence of multiple inputs and outputs makes comparisons difficult. Grey Markovian model has the merits of both grey model and Markov transition probability matrix model. In this paper, DEA and Grey Markovian model are combined to measure the discovery and exploitation goal cost of petroleum enterprise. This method is applied to the cost analysis of Huabei Oil Company of PetroChina, which gives strong support to the feasibility and availability of the study in this paper. [ABSTRACT FROM AUTHOR]

Portfolio selection of oil/ gas projects is a fundamental subject of capital budgeting in the energy sector. In industry, strategic planning requires the selection of a portfolio of projects to advance the corporate goals. The term portfolio selection is more devoted to a set of financial assets that are easily traded and divisible. The principles and fundaments of portfolio selection can be used to compare strategies of investments, so that firms get maximum returns. In this paper, we discuss the oil/gas projects valuation methods first. Considering the characteristic of the oil/gas project, Real options method is more suitable for the oil/gas project than NPV method, and then a new model extended from the Black-Scholes model is built up. Finally, this paper put forward the portfolio selection model and gives a Computational case. [ABSTRACT FROM AUTHOR]

As a benchmark for measuring market risk, Value-at-Risk (VaR) reduces the risk associated with any kind of asset to just a number (amount in terms of a currency), which can be well understood by regulators, board members, and other interested parties. This paper employs a new kind of VaR approach due to Engle and Manganelli [4] to forecasting oil price risk. In doing so, we provide two original contributions: introducing a new exponentially weighted moving average CAViaR model and developing a least squares regression model for multi-period VaR prediction. [ABSTRACT FROM AUTHOR]

Researches on the similarity measure between XML documents are being progressed in order to effectively control and retrieve various XML documents. Previous works mostly suggest similarity-measuring methods focusing only on the tag structure of XML documents. However, they have a problem of incorrectly calculating the semantic similarity of XML contents. In this paper, we propose a new similarity measurement method considering not only the structural information of tags in XML documents but also the semantic information of tags and text content information related with the tags. Our experiments demonstrate that our proposed method improves the accuracy of similarity, compared to the previous works. [ABSTRACT FROM AUTHOR]

Skyline queries, which return the objects that are better than or equal in all dimensions and better in at least one dimension, are useful in many decision making and monitor applications. With the number of dimensions increasing and continuous large volume data arriving, mining the approximate skylines over data stream under control of losing quality is a more meaningful problem. In this paper, firstly, we propose a novel concept, called approximate skyline. Then, an algorithm is developed which prunes the skyline objects within the acceptable difference and adopts correlation coefficient to adjust adaptively approximate query quality. Furthermore, our experiments show that the proposed methods are both efficient and effective. [ABSTRACT FROM AUTHOR]

Recently, a fast increment of spatio-temporal data volume has been achieved and more importantly the data might distribute everywhere. So, there is a need for spatio-temporal data mining systems that are able to support such distributed spatio-temporal query and analysis operations. Distributed spatio-temporal data mining technologies were discussed in this paper. After discussing the process of spatio-temporal data mining in distributed environment, one actual DSTDMS (Distributed Spatio-Temporal Data Mining System) was designed and then implemented. The system is based on data model of sequent snapshot and accomplished through spatio-temporal extension on PostgreSQL. Various spatio-temporal analyses and mining queries could be carried out in the system through simple SQL statements. By using the system, effective mining of distributed spatio-temporal data were achieved. [ABSTRACT FROM AUTHOR]

In this paper, we apply nonlinear support vector machines to identify peer-to-peer (p2p) traffic in high-performance routers with packet sampling. Due to their high port rates, those routers cannot extract the headers of all the packets that traverse them, but only a sample. The results in this paper suggest that nonlinear support vector machines are highly successful and outperform recent approaches like [1,2]. [ABSTRACT FROM AUTHOR]

Clustering streaming time series is a difficult problem. Most traditional algorithms are too inefficient for large amounts of data and outliers in them. In this paper, we propose a new clustering method, which clusters Bi-clipped (CBC) stream data. It contains three phrases, namely, dimensionality reduction through piecewise aggregate approximation (PAA), Bi-clipped process that clipped the real valued series through bisecting the value field, and clustering. Through related experiments, we find that CBC gains higher quality solutions in less time compared with M-clipped method that clipped the real value series through the mean of them, and unclipped methods. This situation is especially distinct when streaming time series contain outliers. [ABSTRACT FROM AUTHOR]

Rapid advancement and new information sources in many academic fields offer expanded opportunities for collaborative teaching. In this paper, we propose a collaborative teaching approach using integrated electronic environments for information security. Taking the security problem of implicit type conversion of C++ program as instance, we present the learning goals & objectives, course content, student assignments and discussion of the collaborative teaching approach using integrated electronic environments. The teaching performances suggest that this new proposed teaching approach is very efficacious. [ABSTRACT FROM AUTHOR]

Question answering service in e-learning environment is an important issue. Chinese semantic analysis is a key bottleneck for question answering service to understand question's content. This paper proposes a question understanding model to understand and process syntactic and semantic structure. In this paper, we analyzed a lot of questions from students, and clustered questions based on knowledge points. The question understanding model is made to get question focus and question type. According to question focus and question type, the question answering service can precisely know the question's answer by locating knowledge point's attribute. This method could more perfectly understand semantic content of questions than using pure Chinese semantic analysis. It is very useful for students to study in a self-learning environment. [ABSTRACT FROM AUTHOR]

In wireless sensor networks, maximizing battery life (network lifetime), and thus the number of messages the networks can support (network capacity), is a key issue. In this paper, we propose an algorithm, called Most Power Reliable Paths (MoPoRePa), for on-line message routing in energy constrained wireless sensor networks. The time complexity of MoPoRePa is significantly lower than the well-known max min zPmin algorithm. Moreover, simulation studies show that MoPoRePa about 14.75 % and 8.05 % improvement over the max min zPmin algorithm in terms of network capacity and the network lifetime, respectively. [ABSTRACT FROM AUTHOR]

The black and white traveling salesman problem (BWTSP) is a new class of NP-hard problem arising from work on airline scheduling and telecommunication fiber networks. The existing Ghiani LP for the undirected BWTSP contains an exponential number of constraints. For a special case of the directed BWTSP whose L = + ∞, the LP with polynomial number of constraints could be obtained by transforming it to an asymmetric traveling salesman problem with replenishment arcs (RATSP), whereas there exists no LP for the directed BWTSP in its general form. This paper proposes a LP with 3n2 + 2n constraints only for the directed BWTSP in such a way that, by reducing the problem to an asymmetric traveling salesman problem (ATSP), we add n2 cardinality constraints and n2 length constraints to the existing Gavish-Grave LP for the ATSP. The new LP is also valid for the undirected BWTSP when viewed as a special case of the directed BWTSP. [ABSTRACT FROM AUTHOR]

This paper presents a framework for Web Service based workflow optimisation in Grid environment. In this framework, parallel service is introduced as adaptor and container of parallel block of workflow. In parallel service, workflow is divided into multi sub-workflows, which are dispatched and executed in parallel at different workflow engines in different grid nodes. By this strategy, the performance of workflow can be enhanced greatly. In this paper, we also give a strategy how to guarantee the correctness in complicated network environment by redundant computing. [ABSTRACT FROM AUTHOR]

In wireless sensor networks(WSNs), as sensor nodes are characterized by having specific requirements such as limited energy availability, low memory and reduced processing power, energy efficiency is a key issue in designing the network routing. In the existing clustering-based routing protocols for the WSNs, the cluster-heads are usually selected randomly and do a lot of work, which may cause unbalanced energy consumption and thus short network lifetime. Based on this approach, the paper proposes a method that a cluster-head distributes energy load evenly among its members based on their energy usage by chain. With the chain, in a cluster, different node is selected as Agent Node to send the data to cluster-head at different cluster cycle. The new method balances the nodes' power depletion. Simulation results show that it has better power management performance and can prolong the lifetime of the network. [ABSTRACT FROM AUTHOR]

Weather forecast is a complex multi-disciplinary problem which requires a cascade of different scientific tools, from differential equation solvers to high-dimensional statistical and data-mining algorithms. The demand for high-resolution predictions is continuously increasing due to the multiple applications in hydrology, agronomy, etc., which require regional meteorological inputs. To fill the gap between the coarse-resolution lattices used by global weather models and the regional needs of applications, a number of statistical downscaling techniques have been proposed. In this paper we describe a Web portal which integrates the necessary tools with Grid middleware allowing for distributed data access and computing. The portal is part of the ENSEMBLES EU-funded project and allows end users to interactively downscale weather predictions using a web browser. Both the architecture and the usage of the portal are described in this paper. [ABSTRACT FROM AUTHOR]

Based on ServiceBSP model, a resource load balancing algorithm with Multi-Agent is put forward in this paper which achieves the goal of dynamic load balancing and favorable fault-tolerant. The algorithm calculates the load value according to the status of usage of resources of a node and scheduling tasks relies on the load value, while updating the load information dynamically depending on Multi-Agent. The method also can avoid frequent communications on load information. Furthermore, the paper introduces the function of agents, relations and communications among agents in details. Finally, by comparing response time and distribution of load using proposed method with other available methods such as without no load balancing and load balancing only based on the usage of CPU, the experimental simulation shows that the load balancing based on Multi-Agent possesses superior performance on response time and load balancing. [ABSTRACT FROM AUTHOR]

This paper proposes HVEM Control System, which is the cornerstone for teleinstrumentation infrastructure. The proposed system is oriented for people whose primary work is to get access to a remotely located instrument, a High-Voltage Electron Microscope (HVEM). Our system is implemented to fulfill all the necessary requirements in allowing the user to 1) control every single part of HVEM in a fine-grained manner, 2) check the HVEM and observe various states of sample, and 3) take their high resolution images of the sample. In that aspect, this paper suggests an HVEM Control System designed on the concept of the Grid and Web Service which satisfies various types of user groups, explains novel characteristics of our system. [ABSTRACT FROM AUTHOR]

Building and maintaining the information superiority is the basis of the principle of active defense, integrated guarding in the cyberspace. In this paper, a novel method based on grey verhulst model was introduced to forecast the network system's security situation. Starting with unequal interval original risk data series, the proposed method choosed grey verhulst model or its inverse function to forecast the future risk value of network system, ant then it could modify the forecasting precision based on multilevel residual error. Simulation results reveal that the presented model not only gains the intuitionistic curve graph of network security situation, but also can achieve satisfactory precision. At the same time, it is simple in use and deserves further study to fully explore its potential for evaluation issues. [ABSTRACT FROM AUTHOR]

Flood disaster risk analysis, the essential of which is to estimate the probability distribution of flood disaster loss, is an important nonstructural measure for flood control and disaster reduction. In the paper, a frequency analysis method of flood disaster loss based on fuzzy mathematics theory of information diffusion with short time series of flood disaster samples is suggested for flood disaster risk analysis. With the support of historic disaster data, the method has been applied to flood disaster risk analysis for Songhua River Basin and turned out to be effective and practical. [ABSTRACT FROM AUTHOR]

This paper utilizes the fuzzy comprehensive evaluation method (FCEM) to evaluate firm's credit sale risk, and provide the credit line and time limit of a credit sale according to the grades of firm's credit sale risk. Finally, we also present an example to validate the feasibility of this method. The result will greatly benefit in enhancing the firm's credit sale risk management. [ABSTRACT FROM AUTHOR]

The rapid growth of the Internet technology has encouraged organizations to protect their information assets. Furthermore, the need for risk analysis has become very important for organizations. However, the existing risk analysis just presents the guidelines that can be used to determine the security measures but do not support how to evaluate the risks quantitatively. Therefore, in this paper, the quantitative risk evaluation model based on the Markov process, especially for the case of interrelated threats, is proposed. In addition, in order to analyze the relationship between threats, the basic analysis method using the covariance and the correlation coefficient is presented. [ABSTRACT FROM AUTHOR]

There have been many research works on epidemic spreading by mathematical or simulation methods. However, to make it practical as a decision support tool for policy maker, it should be combined with business operation of epidemic control. Thus, an alert and response framework covering outbreak detection, policy analysis and policy execution is proposed in this paper. Key technologies for such a framework at modeling, analysis and data collection aspects are discussed. Sensitivity analysis is proposed for model verification and confidence analysis. Critical point identification is another important topic for control policy making by network topology analysis. The importance and feasibility of these technologies are illustrated by the experiment on Beijing SARS (Severe Acute Respiratory Syndrome) data in 2003. [ABSTRACT FROM AUTHOR]

In order to develop highly secure information systems, it is important to make a security specification of the systems, although it requires heavy labor. Thus database technologies have been applied to software engineering and information security engineering for practical reuse of security specifications. However, because the specifications do not have fixed layout, it is difficult to develop a flexible and useful library for the documents with conventional database technologies. Therefore, this paper proposes a security specification library with a schemaless native XML database. Users of the library can directly store and manage security specifications with any layout. Consequently, the library mitigates the labor for making security specifications. [ABSTRACT FROM AUTHOR]

Energy consumption, economic development level, population growth, and climate change and so on all have close relations with environment pollution. Based on the statistical data of Anhui province in China from 1987 to 2004, this paper analyzes the direct and indirect influences of energy consumption, total output value of industry, population, temperature and precipitation on environment pollution. The results show production pattern of industry and consumption of electric power are main reasons of environmental pollution. [ABSTRACT FROM AUTHOR]

We present a color image segmentation scheme based on pixel color clustering. The proposed segmentation is based on mean shift clustering method, which is a powerful tool for the analysis of feature space. However, earlier clustering techniques based on mean shift use single scale over the entire feature space and are not feasible to analyze complex multi-modal feature spaces. In this paper, we present an adaptive mean shift method, in which the local scale information is involved. Actual regions in an image can be obtained by segmentation using the proposed clustering technique. [ABSTRACT FROM AUTHOR]

This paper develops a comprehensive but simple methodology to valuate information technology investment in a nuclear power station based on fuzzy risk analysis and real option approach. By linking the variability of expected payoffs to specific sources of risk factors, this method could help decision makers achieve a more reliable valuation of the target investment under multiple sources of risk, and avoid complex estimation tasks at the same time. The proposed approach can be easily adopted to value any irreversible investment subject to a high degree of uncertainty that is common in the field of energy economics. [ABSTRACT FROM AUTHOR]

A synergetic model (DWT-LSSVM) is presented in this paper. First of all, the raw data is decomposed into approximate coefficients and the detail coefficients at different scales by discrete wavelet transforms (DWT). These coefficients obtained by previous phase are then used for prediction independently using least squares support vector machines (LSSVM). Finally, these predicted coefficients are combined into a final prediction. The proposed model is applied to oil price prediction. The simulation results show that the synergetic model has greater generalization ability and higher accuracy. [ABSTRACT FROM AUTHOR]

As a versatile investment tool in energy markets for speculators and hedgers, the Goldman Sachs Commodity Index (GSCI) futures are quite well known. Therefore, this paper proposes a hybrid model incorporating ARCH family models and ANN model to forecast GSCI futures price. Empirical results show that the hybrid ARCH(1)-M-ANN model is superior to ARIMA, ARCH(1), CGARCH(1,1), EGARCH(1,1) and ARIMA-ANN models on the RMSE, MAPE, Theil IC evaluation criteria. [ABSTRACT FROM AUTHOR]

Support Vector Machines have been a dominant learning technique for almost ten years, moreover they have been applied to supervised learning problems. Recently two-class unsupervised and semi-supervised classification problems based on Bounded C-Support Vector Machines and Bounded ν-Support Vector Machines are relaxed to semi-definite programming [4][11]. In this paper we will present another version to unsupervised and semi-supervised classification problems based on Lagrangian Support Vector Machines, which trained by convex relaxation of the training criterion: find a labelling that yield a maximum margin on the training data. But the problems have difficulty to compute, we will find their semi-definite relaxations that can approximate them well. Experimental results show that our new unsupervised and semi-supervised classification algorithms often obtain almost the same accurate results as the unsupervised and semi-supervised methods [4][11], while considerably faster than them. [ABSTRACT FROM AUTHOR]

Among numerous techniques of information concealment, Spread spectrum watermarking (SSW) has proved to yield improved results when robustness against attack is at a premium. SSW hides information which is called watermark by spreading its spectrum and then adding them to a host image as a watermarked image. Spreading spectrum is done by a pseudo-noise (PN) sequence. But in standard SSW approaches, receiver must use a separate channel to achieve PN sequence used at the transmitter for detecting hidden information. The unique PN sequence must have a low cross correlation and thereby can be duplicated easily by hostile attackers. In this paper a novel approach based on the unconventional Random Encoding Spread Spectrum is proposed for recovering the spreading sequence of watermark signal without any information from the transmitter. It is contributed a higher secure feature by using of the time-varying random encoded spread spectrum. [ABSTRACT FROM AUTHOR]

Many classification methods are used in microarray gene expression data analysis to identify genes that are predictive to clinical outcomes (survival/fatal) of certain diseases. However, the reliability of these methods is often not well established due to the imprecision of the method and uncertainty of the dataset. In this paper, a knowledge-based belief reasoning system (BRS) is proposed to solve the problem by dealing with the uncertainties inherent in the results of various classification methods. Through the belief combination process, we pursue a means to reduce the uncertainty and improve the reliability of classification so that the underlying features of gene behavior recorded in the microarray expression profiles could be convincingly revealed. [ABSTRACT FROM AUTHOR]

The electron microscope is widely used in chemistry, physics and biology fields, especially in biomedicine ultrastructure research. Using the ultrasonic waves, the time that we spend in preparing the biological specimen will be decreased from 5~10 days to 3~5 hours. A few investigation and discussion of fast preparation was proposed in the paper. This work is very significant to biomedicine research and clinic pathology diagnoses aspects. It is better to choose amplitude with low frequency for the purpose of fast preparation. [ABSTRACT FROM AUTHOR]

We introduce a novel method of automatic providing multimedia content in mobile environment. With using our method, the problem of the limitation of resource of mobile devices can be solved. In this paper, we introduce a novel method of recommendation of personalized contents according to preference clones using a collaborative filtering technique in mobile environment. We divide the user group to two sub-groups by analyzing the match of preferences of members in the sub-groups. The division process recursively applies to each sub-group and place the sub-groups in a binary decision tree (BDT). From the binary decision tree, we identify the preference clones of each target user by matching the target user's consumption behavior to that of each sub-group in the BDT with a sequential manner. We also implemented our system based on Java Micro Edition platform. [ABSTRACT FROM AUTHOR]

With the growing importance of XML in data exchange, much research has been done in proving flexible query mechanisms to extract data from structured XML documents. The semi-structured nature of XML data and the requirements on query flexibility pose unique challenges to database indexing methods. Recently, ViST that uses suffix tree and B+Tree was proposed to reduce the search time of the documents. However, it can cause a lot of unnecessary computation and I/O when processing structural join queries because the numbering scheme of ViST is not optimized. In this paper, we propose a novel index structure to solve the problems of ViST. Our index structure provides the bottom-up query processing method to efficiently process structural queries. Our experiments show that the proposed index structure is efficient in processing both single-path and branching queries with various wild-cards (‘*' and ‘//'). [ABSTRACT FROM AUTHOR]

An incremental Content Based Image Retrieval (CBIR) method is proposed in this paper. This method is based on color histogram. Standard histograms, because of their efficiency and insensitivity to small changes, are widely used for content based image retrieval. We define an algorithm that utilizes the concept of Histogram Refinement [1] and we call it Color Refinement Method. Color Refinement method splits the pixels in a given bucket into several classes just like histogram refinement method. The classes are all related to colors and are based on color coherence vectors. After the calculation of clusters using color refinement method, inherent features of each of cluster is calculated. These inherent features are used for incremental CBIR. [ABSTRACT FROM AUTHOR]

Internet search engine techniques have evolved from simple web searching using categorization (e.g., Yahoo) to advanced page ranking algorithms (e.g., Google). However, the challenge for the next generation of search algorithms is not the quantity of search results, but identifying the most relevant pages based on a semantic understanding of user requirements. This notion of relevance is closely tied to the semantics associated with the term being searched. The ideal situation would be to represent results in an intuitive way that allows the user to view their search results in terms of concepts related to their search word or phrase rather than a list of ranked web pages. In this paper, we propose a semantic clustering approach that can be used to build a conceptual search engine. [ABSTRACT FROM AUTHOR]

Software document repositories store artifacts produced in the course of developing software products. But most repositories are simply archives of documents. It is not unusual to find projects where different software artifacts are scattered in unrelated repositories with varying levels of granularity and without a centralized management system. This makes the information available in existing repositories difficult to reuse. In this paper, a methodology for constructing an ontology-based repository of reusable knowledge is presented. The information in the repository is extracted from specification documents using text mining. Ontologies are used to guide the extraction process and organize the extracted information. The methodology is being used to develop a repository of recurring and crosscutting aspects in software specification documents. [ABSTRACT FROM AUTHOR]

Keeping people away from litigious information becomes one of the most important research area in network information security. Indeed, Web filtering is used to prevent access to undesirable Web pages. In this paper we review some existing solutions, then we propose a violent Web content detection and filtering system called "WebAngels filter" which uses textual and structural analysis. "WebAngels filter" has the advantage of combining several data-mining algorithms for Web site classification. We discuss how the combination learning based methods can improve filtering performances. Our preliminary results show that it can detect and filter violent content effectively. [ABSTRACT FROM AUTHOR]

This paper discusses a framework for image retrieval. Most current systems are based on a single technique for feature extraction and similarity search. Each technique has its advantages and drawbacks concerning the result quality. Usually they cover one or two certain features of the image, e.g. histograms or shape information. The proposed framework is designed to be highly flexible, even if performance may suffer. The aim is to give people a platform to implement almost any kind of retrieval issues very quickly, whether it is content based or somehing else. The second advantage of the framework is the possibility to change retrieval characteristics within the program completely. This allows users to configure the ranking process as needed. [ABSTRACT FROM AUTHOR]

Data mining is the process of automatically finding implicit, previously unknown and potentially useful information from large volumes of data. Embedded systems are increasingly used for sophisticated data mining algorithms to make intelligent decisions while storing and analyzing data. Since data mining applications are designed and implemented considering the resources available on a conventional computing platform, their performance degrades when executed on an embedded system. In this paper, we analyze the bottlenecks faced in implementing these algorithms in an embedded environment and explore their portability to the embedded systems domain. Particularly, we analyze the floating point computation in these applications and convert them into fixed point operations. Our results reveal that the execution time of five representative applications can be reduced by as much as 11.5× and 5.2× on average, without a significant impact on accuracy. [ABSTRACT FROM AUTHOR]