Most Read articles

Published in last 1 year |  In last 2 years |  In last 3 years |  All

All

Please wait a minute...


For Selected: Download Citations EndNote Ris BibTeX Toggle Thumbnails

Select

Super-Resolution Image Reconstruction with Iterative Back Projection Algorithm GUO Weiwei 1, ZHANG Pinzheng 2+ Journal of Frontiers of Computer Science and Technology    2009, 3 (3): 321-329.   DOI: 10.3778/j.issn.1673-9418.2009.03.010 An image super-resolution reconstruction method， which combines the frequency domain motion estimation and iterative back projection （IBP） algorithm， has been proposed. Based on the frequency domain phase difference among the low resolution images， the motion parameters between the reference LR image and other LR images are estimated. Combining the IBP algorithm and the estimated motion parameters， the super-resolution image is iteratively reconstructed. Experimental results demonstrate that the proposed method is effective. Related Articles | Metrics

Abstract （11656）

PDF （4282）

Select

A Survey on Question and Answering Systems MAO Xianling, LI Xiaoming Journal of Frontiers of Computer Science and Technology    2012, 6 (3): 193-207.   DOI: 10.3778/j.issn.1673-9418.2012.03.001 Recently, question and answering systems have attracted lots of attention. Given a question, the goal of question and answering is to return a concise, exact answer. According to the format of data, question and answering can be di-vided into three categories: the structural data based question and answering, the free-text based question and answering, the question-answer pairs based question and answering. This paper describes and summarizes the char-acteristics and related researches of these three categories respectively. Then, it discusses the future work of question and answering. Related Articles | Metrics

Abstract （11563）

PDF （12062）

Select

Quantum computing and quantum computers WU Nan,SONG Fang-min+ Journal of Frontiers of Computer Science and Technology    2007, 1 (1): 1-16.   As a new computing model，quantum computing can in principle exploit quantum mechanical effects to perform computations more rapidly than classical Turing Machine model based on the principles of quantum computing and some important quantum.Meanwhile，the physical implementations of quantum computer are also proposed.Optic photon quantum computer and quantum computer based on Nuclear Magnetic Resonance（NMR），ion traps or harmonic oscillator，etc.，have been realized one by one.Several quantum algorithms appear in recent years.A successful demonstration of Shor’s quantum algorithm is made by using a quantum computing device based on NMR in 2001，and it shows the quantum computer’s great potential power for processing complicated problems.Finally，the developments in the aspect of physical implementation of quantum computer are summarized． Related Articles | Metrics

Abstract （9390）

PDF （7304）

Select

Improved Algorithm Based on Girvan-Newman Algorithm for Community Detection ZHU Xiaohu 1,2, SONG Wenjun 1,2, WANG Chongjun 1,2+, XIE Junyuan 1,2 Journal of Frontiers of Computer Science and Technology    2010, 4 (12): 1101-1108.   DOI: 10.3778/j.issn.1673-9418.2010.12.004 To improve the efficiency of Girvan-Newman(G-N) algorithm, a community detection algorithm named modularity extreme approximation (MEA) is given. MEA algorithm uses the increment of modularity as the meas-ure for community structure and finds the solution with a greedy strategy. The theoretical analysis and experimental results show the MEA algorithm is more effective and faster than the G-N algorithm. Related Articles | Metrics

Abstract （9221）

PDF （2876）

Select

Software Engineering Data Mining: A Survey YU Shusi, ZHOU Shuigeng, GUAN Jihong Journal of Frontiers of Computer Science and Technology    2012, 6 (1): 1-31.   DOI: 10.3778/j.issn.1673-9418.2012.01.001 With the rapid enlargement of software scale, to retrieve manually the relevant information of software development and maintenance is becoming more and more difficult. Data mining technology can help to discover useful information from software engineering data automatically, which thus speeds up the process of software de-velopment. This paper surveys the state of the art techniques of software engineering data mining. First, it presents basic concepts and technical challenges of software engineering data mining. Then, it discusses the details of data mining at different phases of software engineering, including motivation, problems, procedures and approaches, specifically, it emphasizes the methods of data pre-processing and representation. Finally, it gives a vision of future development of software engineering data mining technology. Related Articles | Metrics

Abstract （9128）

PDF （4429）

Select

Survey of Metamorphic Testing DONG Guowei 1,2, XU Baowen 1,2+, CHEN Lin 1,2, NIE Changhai 1,2, WANG Lulu 1,2 Journal of Frontiers of Computer Science and Technology    2009, 3 (2): 130-143.   DOI: 10.3778/j.issn.1673-9418.2009.02.002 Metamorphic testing （MT） is a technique that tests software by checking relationships among several executions， so it is very practical and effective for program with oracle problem. This technique has been used in various fields， and great developments have been made in MT process optimization and combination with other testing methods since 1998. The current investigation of MT is surveyed， and some research directions for it are presented， which are the research of MT sufficiency， the effective metamorphic relation constructing technology， the effective original testing case selecting technology， the research of MT for new type software and the development of MT tools， etc. Related Articles | Metrics

Abstract （8810）

PDF （4069）

Select

TF-IDF Similarity Based Method for Tag Clustering HAN Min， TANG Changjie+， DUAN Lei， LI Chuan， GONG Jie Journal of Frontiers of Computer Science and Technology    2010, 4 (3): 240-246.   DOI: 10.3778/j.issn.1673-9418.2010.03.006 As a new concept of Web 2.0， social tagging system aims at expressing users’ interests clearly and specifically. Tag clustering is an important research topic in social tagging system mining. Evaluation similarity among social tags is the key technique in tag clustering. The main contributions include： （1） introduce a new method to calculate the tag similarity based on TF-IDF， and propose a clustering algorithm based on the new method； （2） analyze the conditions that influence tag similarity； （3） conduct extensive experiments to demonstrate that proposed method is more efficient compared with some methods proposed before. Related Articles | Metrics

Abstract （8748）

PDF （1845）

Select

Research and Optimization of Nutch Distributed Crawler ZHAN Hengfei 1+, YANG Yuexiang 2, FANG Hong 2 Journal of Frontiers of Computer Science and Technology    2011, 5 (1): 68-74.   DOI: 10.3778/j.issn.1673-9418.2011.01.007 As a good open-source search engine, Nutch kernel code uses a lot of MapReduce programming models,being used by more and more businesses and organizations to customize their needs in line with the distributed search engine product. As a good search engine, one of the important prerequisites is how to grab network data as much as possible to build indexes. This paper introduces Nutch’s working mechanism based on Hadoop distributed Web crawler, points out its shortcomings and proposes an improved program, which can make Web crawler using network resources more efficiently to capture network data. Experimental results show that it is indeed more efficient than the original programs. Related Articles | Metrics

Abstract （8720）

PDF （3781）

Select

Programming the VFDT Algorithm in Data Stream Manage-ment System* YUAN Lei 1, ZHANG Yang 2+, LI Mei 1, LI Xue 3, WANG Yong 4 Journal of Frontiers of Computer Science and Technology    2010, 4 (8): 673-682.   DOI: 10.3778/j.issn.1673-9418.2010.08.001 Integrating data stream mining algorithm with data stream management system (DSMS) is a novel challenge for data mining and database researchers. But the integration of very fast decision tree (VFDT) with data stream management has not been reported till now. This paper focuses on integrating VFDT algorithm with Esper by exploiting capabilities of data stream management system (DSMS). How to transform the algorithm into efficient Esper query language (EQL) is analyzed, and two implementations for integrating the popular VFDT algorithm with DSMS are proposed: Transforming the VFDT algorithm into EQL straightforwardly (denoted by DVFDT); an opti-mized version of DVFDT based on the inherent batch mode of Esper (denoted by optimal-DVFDT). The proposed implementations with VFDT based on Java (denoted by JVFDT) in terms of classification accuracy and performance are compared. Experiments on a set of large volume of synthetic data show the implementation works efficiently and accurately. In addition, this approach also has better performance for the sub-streams of the original data stream. Related Articles | Metrics

Abstract （8632）

PDF （2116）

Select

A new hierarchical clustering algorithm based on tree edit distance QIAO Shaojie 1,2,TANG Changjie 1+,CHEN Yu 1,PENG Jing 3,WEN Fenlian 1 Journal of Frontiers of Computer Science and Technology    2007, 1 (3): 282-292.   In order to recognize the false status which has been forged and tempered by suspects,a new method is proposed to compute attribute similarities based on tree edit distance,and its mathematical properties are proved. The paper proposes a new clustering algorithm based on hierarchical encoding method named HCTED（Hierarchical Clustering Algorithm Based on Tree Edit Distance）. This method uses tree edit distance to compute attribute similarities with minimum cost,overcomes the shortage of traditional clustering algorithms and improves the precision of clustering according to the predefined threshold. Experiments demonstrate that the new method is accurate and efficient in identity recognition,discuss the effects of different experimental parameters,and show that HCTED is more accurate and faster than traditional clustering algorithms. The new algorithm has been used in data analysis of transient population for public security successfully. Related Articles | Metrics

Abstract （8608）

PDF （2485）

Select

Depth Camera in Computer Vision and Computer Graphics: An Overview XIANG Xueqin, PAN Zhigeng, TONG Jing Journal of Frontiers of Computer Science and Technology    2011, 5 (6): 481-492.   An increasing number of applications depend on accurate and fast 3D scene analysis, such as geometry reconstruction, collision prevention, mixed reality, and gesture recognition. The acquisition of a range map by image- based analysis or laser scan techniques is still time-consuming and expensive. Emerged as an alternative device to measure distance, depth camera enjoys some advances, e.g., lower price and higher photo speed, that have not be made in traditional 3D measuring systems. Recently, significant improvements have been made in order to achieve low-cost and compact depth camera devices that have the potential to revolutionize many fields of research, including computer vision, computer graphics and human computer interaction (HCI). These technologies are also starting to attract many researchers working for academic or commercial purposes. This paper gives an overview of recent developments in depth camera technology and discusses the current state of the integration of this technology into various related applications in computer vision and computer graphics. Related Articles | Metrics

Abstract （8293）

PDF （4595）

Select

Survey of location privacy-preserving PAN Xiao+,XIAO Zhen,MENG Xiaofeng Journal of Frontiers of Computer Science and Technology    2007, 1 (3): 268-281.   With blooming of sensor and wireless mobile devices,it is easy to access mobile users’ location information anytime and anywhere. On one hand,Location Based Services（LBS） is becoming more and more valuable and important. On the other hand,location privacy issues raised by such applications have also grasped more and more attention. However,due to the specificity of location information,traditional privacy-preserving techniques in data publish can not be used. The paper analyzes the challenges of location privacy-preserving,and gives a survey of existing work including the system architecture,location anonymity and query processing. Finally,some open issues are given at the end of the paper. Related Articles | Metrics

Abstract （8184）

PDF （6213）

Select

Review of medical image processing and analyzing software platforms LI Enzhong+ Journal of Frontiers of Computer Science and Technology    2008, 2 (5): 467-477.   DOI: 10.3778/j.issn.1673-9418.2008.05.002 The technology of medical image processing and analyzing can provide more clear and accurate images and more useful quantitative information for the doctor to assist the diagnosis and treatment. The development of algorithms such as segmentation， registration and 3D visualization is the motivation of medical image processing and analyzing， while the software platforms based on these algorithms are the accelerators of medical image processing and analyzing. Many software platforms have been designed for scientific research or providing wieldy assistant tool for certain users. This paper introduces some mainstream platforms， and discusses the advantages and shortcomings of these platforms adequately. Our aim is to spread the applications of medical image processing and analyzing software platform， and motivate the researchers to contribute to the development of the software platforms. Related Articles | Metrics

Abstract （8149）

PDF （6511）

Select

Skyline Computation under MapReduce Framework ZHANG Boliang 1,2, ZHOU Shuigeng 1,2, GUAN Jihong 3 Journal of Frontiers of Computer Science and Technology    2011, 5 (5): 385-397.   Skyline computation, due to its wide applications in multi-objective decision making and data visualization, has attracted many research interests in database community recently. Aiming at cloud computing applications, this paper addresses the problem of Skyline computation under the MapReduce framework. As a parallel programming model for data-intensive computing applications, MapReduce runs on a cluster of commercial PCs with the main idea of task decomposition and solution reduction. Based on different data division strategies, this paper proposes three algorithms: MapReduce based block-nested-loops (MR-BNL), MapReduce based sort-filter-skyline (MR-SFS) and MapReduce based bitmap (MR-Bitmap). It conducts extensive experiments to evaluate and compare the three algorithms under different situations of different data distributions, dimensions and buffer sizes. Related Articles | Metrics

Abstract （8053）

PDF （2542）

Select

Survey of Information-Centric Networking XIA Chunmei, XU Mingwei Journal of Frontiers of Computer Science and Technology    2013, 7 (6): 481-493.   DOI: 10.3778/j.issn.1673-9418.1303025 Network requirements are evolved from data communications between two hosts to accessing a large amount of information through networks. The current Internet uses address-centric network communication model, which is suitable for the communications between hosts but not efficient in the communications between hosts and networks. Given this fact, information-centric networking (ICN) becomes the hot spot to meet the new requirements. Firstly, this paper presents the basic idea of ICN, analyzes the features of ICN, and classifies different ICN approaches, including the following aspects: naming, resolving, routing and caching. Secondly, this paper summarizes the five key technologies in ICN which are naming, resolving, routing, distribution and caching, and then analyzes and evaluates the existing ICN schemes. Finally, several key issues in ICN are discussed and future directions are pointed out. Related Articles | Metrics

Abstract （8017）

PDF （5876）

Select

Research on Kmeans Algorithm Optimization Based on OpenCL WU Zailong, ZHANG Yunquan, XU Jianliang, JIA Haipeng, YAN Shengen, XIE Qingchun Journal of Frontiers of Computer Science and Technology    2014, 8 (10): 1162-1176.   DOI: 10.3778/j.issn.1673-9418.1312042 As a typical clustering algorithm and an important method to data decomposition and packet processing, Kmeans algorithm is widely used in image processing, machine learning and biology, etc. Due to the constant expansion on data set, Kmeans is facing more and more demand on its performance. Having taken into full account the difference between hardware platforms and architectures, this paper conducts a systematic research on achieving Kmeans algorithm efficiently running on GPU and APU platforms based on OpenCL. And with the help of several optimization methods, such as the implementation of iterative algorithm with multiple global synchronization in GPU, the reduction on global synchronization by redundant computation, the redistribution on thread task, the reuse of local memory, etc, Kmeans algorithm achieves high efficient implementation on different hardware architectures and the optimization methods suitable for iterative algorithm are summed up. The experimental results show that the optimized algorithm gets 136.975~170.333 times speedup on AMD HD7970 GPU than the CPU version (with considering the data transfer time) and gets 22.2365~24.3865 times speedup on AMD A10-5800K APU than the CPU version, which effectively verifies the validity and the cross-platform ability of the optimization methods proposed in this paper. Related Articles | Metrics

Abstract （7729）

PDF （2167）

Select

Research on Service-Oriented Data Mining Engine Based on Cloud Computing YU Yonghong, XIANG Xiaojun, GAO Yang, SHANG Lin, YANG Yubin Journal of Frontiers of Computer Science and Technology    2012, 6 (1): 46-57.   DOI: 10.3778/j.issn.1673-9418.2012.01.003 The scalability of data mining algorithms is restricted when dealing with large-scale data. There are significant differences in a wide range of application areas and requirements for knowledge discovery process. It is fundamental to provide effective formalisms to design distributed data mining application and support their efficient execution. This paper proposes a novel service-oriented data minging engine based on cloud computing framework, which is named as CloudDM. Differentiating from grid-based distributed data mining framework, CloudDM exploits the capacity of open source cloud computing platform—Hadoop for large-scale data analysis, supports the design and execution of distributed data mining applications according to SOA (service-oriented architecture). Moreover, it discusses and reports the key component functions and implementation technologies. According to the design principles of SOA and data mining engine based on cloud computing, the paper can solve the problems in massive data mining systems, such as big data storage, data processing and interactive operation of algorithms, etc. Related Articles | Metrics

Abstract （7692）

PDF （1868）

Select

EECS： an energy-efficient clustering scheme in wireless sensor networks CHEN Guihai 1+,LI Chengfa 1,YE Mao 1,WU Jie 2 Journal of Frontiers of Computer Science and Technology    2007, 1 (2): 170-179.   Clustering is an effective topology control approach which can increase network scalability and lifetime in Wireless Sensor Networks（WSNs）. LEACH is an elegant clustering protocol that prolongs the network lifetime. The paper proposes a novel clustering scheme EECS for WSNs， which better suits the periodic data gathering applications. In the cluster head election phase it selects a small portion of nodes to join the election， adopts local radio communications method without iteration， and always selects cluster heads with more residual energy. Further， in the cluster set-up phase it uses a novel method for balancing the load of cluster heads. Its control message overhead is small， and the cluster head distribution is near uniform. EECS also has a high energy effective consuming rate across the network. Simulation results show that EECS outperforms LEACH more than 35% in prolonging the network lifetime under the same assumptions. Related Articles | Metrics

Abstract （7676）

PDF （2543）

Select

Survey of combinatorial test generation WANG Ziyuan 1,2+, XU Baowen 1,2, NIE Changhai 1,2 Journal of Frontiers of Computer Science and Technology    2008, 2 (6): 571-588.   DOI: 10.3778/j.issn.1673-9418.2008.06.002 As a practical software testing approach， Combinatorial Testing （CT） aims to detect the faults that triggered by interactions among factors （or parameters） in SUT by designing and executing a small combinatorial test suite to cover the required combinations of these factors. Up to date， many works have been done in the field of CT， and one of the most important aspect is combinatorial test generation， which aims to generate a small test suite to satisfy a given combinatorial coverage criteria. By describing some representative combinatorial test generation algorithms for two different combinatorial coverage criteria （N-way combinatorial coverage and variable strength combinatorial coverage）， the existing works on combinatorial test generation in this survey are reviewed. Furthermore， combinatorial test generation techniques for prioritization， constraint， and fault location are also presented. At last， the problems of existing works are analyzed， and some future trends in the field of CT are proposed. Related Articles | Metrics

Abstract （7659）

PDF （3420）

Select

S-SimRank： Combining Content and Link Information to Cluster Papers Effectively and Efficiently CAI Yuanzhe1,2, LI Pei1,2, LIU Hongyan3, HE Jun1,2+, DU Xiaoyong1,2 Journal of Frontiers of Computer Science and Technology    2009, 3 (4): 378-391.   DOI: 10.3778/j.issn.1673-9418.2009.04.005 Content analysis and link analysis among documents are two common methods in recommending system. Compared with content analysis， link analysis can discover more implicit relationship between documents. At the same time， because of the noise， these methods can’t gain precise result. To solve this problem， a new algorithm， S-SimRank （Star-SimRank）， is proposed to effectively combine content analysis and link analysis to improve the accuracy of similarity calculation. The experimental results for the ACM data set show that S-SimRank outperforms other algorithms. In the end， the mathematic prove for the convergence of S-SimRank is given. Related Articles | Metrics

Abstract （7648）

PDF （1728）

Select

How to get effective slide-window size in time series similarity search LI Feng 1+, XIAO Jianhua 2 Journal of Frontiers of Computer Science and Technology    2009, 3 (1): 105-112.   DOI: 10.3778/j.issn.1673-9418.2009.01.010 As a non-trivial problem， the most promising solutions of similarity search in time series databases involve performing dimensionality reduction on the original data. The key to this question is how to effectively keep the original time series information while reducing computational complexity at the best effort. It discusses the slide-window practical application in this area， and then concludes how to get an effective slide-window size in similarity of time series research. To find some useful eigenvalue in time series by mining the distributing values of original series， and deduce a collection of candidate slide-window size， find the optimal slide window size by adapting to the series eigenvalue variable motion movement. Related Articles | Metrics

Abstract （7622）

PDF （2675）

Select

Privacy-Preserving Query Processing in Cloud Computing HUO Zheng, MENG Xiaofeng, XU Jianliang Journal of Frontiers of Computer Science and Technology    2012, 6 (5): 385-396.   DOI: 10.3778/j.issn.1673-9418.2012.05.001  A vital concern in cloud computing is how to protect both data privacy and query privacy while providing query services for users. This paper surveys several critical techniques of privacy-preserving query processing in cloud computing, which include cloud database indexing, query optimization, encryption-based privacy-preserving techniques, privacy-preserving techniques based on secure multi-party computation and authorization auditing techniques. Finally, the paper analyzes the challenges of privacy-preserving query processing in cloud computing and figures out the trend of this area. Related Articles | Metrics

Abstract （7604）

PDF （3169）

Select

Proof of equivalence between semantic expressions of derived relational algebra operators YANG Bo 1,2,3+, XUE Jinyun 1 Journal of Frontiers of Computer Science and Technology    2008, 2 (1): 97-103.   The derived relational algebra operators are widely used in the relational database query languages.There are two kinds of representations for their semantics. One is the expression described by original relational operators， the other is the expression described by one order logic. However， the strict equivalence proof of the two different semantic expressions is not given in related documents. The present study manages to prove the equivalence of the two different kinds of semantic expressions through a series of formal deductions step by step according to some related characteristics of relational algebra. The beginning of these deductions is the semantic expression described by original relational operators； the end is the semantic expression described by one order logic. The foundation of correctness proof of relational algebra expressions can be laid by the present deduction method. Related Articles | Metrics

Abstract （7593）

PDF （1861）

Select

GPU-based Parallel SVM Algorithm DO Thanh-Nghi 1, NGUYEN Van-Hoa 2, POULET François 3+ Journal of Frontiers of Computer Science and Technology    2009, 3 (4): 368-377.   DOI: 10.3778/j.issn.1673-9418.2009.04.004 A new parallel and incremental support vector machine （SVM） algorithm for the classification of very large datasets on graphics processing units （GPUs） is presented. SVM and kernel related methods have shown to build accurate models but the learning task usually needs a quadratic program so that this task for large datasets requires large memory capacity and long time. A recent least squares SVM （LS-SVM） proposed by Suykens and Vandewalle for building incremental and parallel algorithm is extended. The new algorithm uses graphics processors to gain high performance at low cost. Numerical test results on UCI and Delve dataset repositories show that this para-llel incremental algorithm using GPUs is about 130 times faster than its CPU implementation and often significantly faster （over 2 500 times） than state-of-the-art algorithms like LibSVM， SVM-perf and CB-SVM. Related Articles | Metrics

Abstract （7527）

PDF （2258）

Select

Improved SMOTE Algorithm for Imbalanced Datasets WANG Chaoxue, ZHANG Tao, MA Chunsen Journal of Frontiers of Computer Science and Technology    2014, 8 (6): 727-734.   DOI: 10.3778/j.issn.1673-9418.1403003 Based on analyzing the shortages of SMOTE (synthetic minority over-sampling technique) in the synthesis of minority class samples, this paper presents an improved SMOTE (GA-SMOTE). The key of GA-SMOTE lies on leading three basic genetic operators of genetic algorithm (GA) into SMOTE, making use of the selection operator to achieve the different samples from the minority class and depending on crossover operator and mutation operator to realize the fine control of the synthesis quality to the minority class samples. GA-SMOTE and SVM (support vector machine) are combined to handle the classification problem on imbalanced datasets. A large amount of experiments on the UCI datasets show that GA-SMOTE promises prominent synthesis effect to the minority class samples, and brings better classification performance on imbalanced datasets with SVM. Related Articles | Metrics

Abstract （7439）

PDF （1987）

Select

Survey of Evolutionary Testing XIE Xiaoyuan 1, XU Lei 1, XU Baowen 1,2+,NIE Changhai 1, SHI Liang 3 Journal of Frontiers of Computer Science and Technology    2008, 2 (5): 449-466.   DOI: 10.3778/j.issn.1673-9418.2008.05.001 Evolutionary Testing （ET）， as a very promising technique for automatic testing， can generate effec-tive test cases for many test objects successfully. ET converts the task of test case generation into an optimization problem， and uses Genetic Algorithm（GA） to search for the desired solutions with very little cost. The input domain of system under test is the search space of GA. With a complete automation， this search process can be more efficient. Besides as a result of both the guidance and the randomicity， the search can overcome the shortcoming of complex search domain and avoid a blind search of simple random testing. In recent years， more and more researches have been focused on ET， and have made it being applied successfully in many areas. This paper introduces the application of ET in structural testing， functional testing， temporal performance testing and object-oriented testing. From summarizing the main ET methods in those applications， it analyzes and compares each method’s advantages and shortcomings， in order to provide some valuable advices for studying and selecting proper ET methods. Additionally the paper concludes the performance optimization techniques of ET， and finally it presents ET’s prospect and the future research for ET. Related Articles | Metrics

Abstract （7394）

PDF （4178）

Select

An Ontology Based Dynamic Context Model WU Zhanqing 1,2+, TAO Xianping 1,2, LV Jian 1,2 Journal of Frontiers of Computer Science and Technology    2008, 2 (4): 356-367.   DOI: 10.3778/j.issn.1673-9418.2008.04.003 Context modeling is the essential task to build context-aware systems， it’s the key for applications to understand and utilize context. Considering the lifecycle of context， it can be divided into persistent context and dynamic context； the latter can be further divided into state context and transition context. Many researchers have done lots of works about persistent context and state context modeling， but there are few works dealing with transition context. In this paper， a formal dynamic context model based on ontology using OWL is proposed， as well as a context fusion and consumption mechanism implementation. Also， some programming principles are provided for applications and a system called Auto-Lock Door is demonstrated to show how systems get benefit from the model. Related Articles | Metrics

Abstract （7283）

PDF （1991）

Select

Brief Introduction to SMT Solving JIN Jiwei, MA Feifei, ZHANG Jian Journal of Frontiers of Computer Science and Technology    2015, 9 (7): 769-780.   DOI: 10.3778/j.issn.1673-9418.1405041 SMT is the problem of deciding the satisfiability of a first order formula with respect to some theory formulas. It is being recognized as increasingly important due to its applications in different communities, in particular in formal verification, program analysis and software testing. This paper provides a brief overview of SMT and its theories. Then this paper introduces some approaches aiming to improve the efficiency of SMT solving, including eager and lazy approaches and optimum technique which have been proposed in the last ten years. This paper also introduces some state-of-the-art SMT solvers, including Z3, Yices and CVC3/CVC4. Finally, this paper gives a preliminary prospect on the research trends of SMT, which include SMT with quantifier, optimization problems subject to SMT and volume computation for SMT. Related Articles | Metrics

Abstract （7276）

PDF （7019）

Select

Survey of Test Case Prioritization for Regression Testing QU Bo 1+, NIE Changhai 2,3, XU Baowen 2,3 Journal of Frontiers of Computer Science and Technology    2009, 3 (3): 225-233.   DOI: 10.3778/j.issn.1673-9418.2009.03.001 Test case prioritization is an effective and practical technique applied in regression testing. It helps to increase the effectiveness of test suite at meeting some performance goals by sorting and executing test cases according to some criterion. Firstly， the background and basic concept of test case prioritization are introduced. And then， the existing research work is concluded and the performance of prioritization approaches in different context is compared. Finally， some unsolved problems are presented and the future work about test suite prioritization is given. Related Articles | Metrics

Abstract （7246）

PDF （3012）

Select

High Effective Two-round Remote File Fast Synchronization Algorithm XU Dan1+, SHENG Yonghong2, JU Dapeng2, WU Jianping1, WANG Dongsheng2,3 Journal of Frontiers of Computer Science and Technology    2011, 5 (1): 38-49.   DOI: 10.3778/j.issn.1673-9418.2011.01.004 Fast remote file synchronization has a widespread application in many scenarios such as the file backup and recovery, Web and ftp site mirroring, content distribution network, Web access and so on. This paper presents a high effective two-round fast synchronization algorithm tpsync which combines content-based variable-sized chunk and fixed-sized sliding block methods. tpsync is implemented with two rounds. For the first round, tpsync adopts content-based variable-sized chunk to locate the local change between similar files in coarse-grained scale. In the second round, tpsync looks up the differential data in the local changed data segment with fixed-sized sliding block method in fine-grained scale, and finally achieves the file synchronization by two-round data interaction. This paper executes a comparison experiment between tpsync and the traditional single-round synchronization method rsync.Extensive experiments on text, binary and database files demonstrate that tpsync can achieve a higher performance on average synchronization time and the amount of network traffic data than rsync. Related Articles | Metrics

Abstract （7223）

Select

A survey of Low-rate Denial-of-Service attacks HE Yanxiang 1,2, LIU Tao 1+, CAO Qiang 1, XIONG Qi 1, HAN Yi 1 Journal of Frontiers of Computer Science and Technology    2008, 2 (1): 1-19.   Low-rate denial-of-service attack is a novel category of attacks that are aimed at exploiting the adaptive behavior exhibited by several network and system protocols. Different from traditional DoS attacks， this kind of attacks can make serious destroy on the victims by using periodically non-suspicious low-rate attack streams. Since they have been brought forward， these new attacks have caused special concern； the problems of detecting and defending towards them have gradually become important research issues in the network security area. In the paper， theoretical analyses， modeling and simulations of various LDoS attacks are presented， the difficulties of defending and current solutions are discussed. At the end of the paper， several problems which need further researches are put forward， in order to provide reference to the future research work on the defending of this kind of attacks. Related Articles | Metrics

Abstract （7199）

PDF （4590）

Select

RFID complex event processing techniques GU Yu,YU Ge+,ZHANG Tiancheng Journal of Frontiers of Computer Science and Technology    2007, 1 (3): 255-267.   With the development of RFID technology,RFID is being applied ubiquitously in every field. With throughly processing and analyzing RFID data,more complex events and implicit knowledge could be discovered to effectively support advanced applications such as event monitoring and pre-warning. Due to the speciality of RFID,it is insufficient to take use of the state of art active database and data stream techniques for high-performance RFID event detection and processing. The paper analyzes the characteristics of RFID data,makes a survey of the newest technology over RFID complex event processing,and explores the new problems to be solved urgently,including RFID data cleaning,data-centric event detection techniques,event-centric event detection techniques and complex event processing systems,and finally gives an outlook of the research issues in the future. Related Articles | Metrics

Abstract （7191）

PDF （4199）

Select

CommTracker： A Core-based Algorithm of Tracking Community Evolution WANG Yi+, WU Bin, YANG Shengqi Journal of Frontiers of Computer Science and Technology    2009, 3 (3): 282-292.   DOI: 10.3778/j.issn.1673-9418.2009.03.006 CommTracker， a novel and parameter-free algorithm of tracking community evolution is proposed， which utilizes the representative quality of core nodes in a community to establish the evolving relationship between two communities in consecutive time snapshots. With such a distinct strategy， it is suitable for analyzing large scale datasets. Depending on relationships established from CommTracker， it is feasible to identify community split and mergence. In addition， two relationships amongst evolution traces， evolution traces intersection and community rebirth， are also studied. At last， the correctness and effectiveness of our algorithm on 4 real datasets are demonstrated. Related Articles | Metrics

Abstract （7066）

PDF （1769）

Select

PCP theorem and its applications to research on non-approximatable problems XU Daoyun + Journal of Frontiers of Computer Science and Technology    2008, 2 (1): 20-31.   The PCP theorem is one of important results in complexity for recent ten years. The paper introduces the evolution from Turing computation model to the Probabilistically Checkable Proofs（PCP）， the basic theory of PCP and its principle and approach to non-approximatable problems. Related Articles | Metrics

Abstract （7009）

PDF （3541）

Select

Identification of Plant Resistance Gene with Random Forest GUO Yingjie, LIU Xiaoyan, GUO Maozu, ZOU Quan Journal of Frontiers of Computer Science and Technology    2012, 6 (1): 67-77.   DOI: 10.3778/j.issn.1673-9418.2012.01.005 The traditional homology sequence alignment based approaches usually have high false positive rate and consequently new resistance genes are difficult to be identified. This paper presents a resistance gene identification approach by applying random forest classifier and K-Means under-sampling method. In order to solve the aimless problem in gene-mining research, two main contributions are provided. Firstly, it introduces random forest and 188 dimension features to identify resistance genes, accordingly the sample statistic learning approach can efficiently capture the internal characteristic of resistance genes. Secondly, it selects a more representative training subset and reduces the identification errors for solving the serious imbalanced classification during the training process. The experimental results indicate that the approach can efficiently identify the resistance genes, not only precisely clas-sifying the existing experimental verified data, but also obtaining high accuracy on the negative sample dataset. Related Articles | Metrics

Abstract （6992）

PDF （1738）

Select

Matching Method of Remote Sensing Images Based on SURF Algorithm and RANSAC Algorithm CHEN Yixia, SUN Quansen, XU Huanyu, GENG Leilei Journal of Frontiers of Computer Science and Technology    2012, 6 (9): 822-828.   DOI: 10.3778/j.issn.1673-9418.2012.09.006 This paper proposes a matching method for remote sensing images, which combines the superiorities of the speeded up robust features (SURF) algorithm and the random sample consensus (RANSAC) algorithm. Firstly, feature detection and pre-matching of images are done by using SURF algorithm. Secondly, the mismatching is wiped out by using RANSAC algorithm. This method solves the mismatching problem of image matching. Integrated experiments on feature detection and matching as well as the settlement of transformation matrix show that the proposed method is effective. Related Articles | Metrics

Abstract （6942）

PDF （2049）

Select

Application of Complex Networks Clustering Algorithm in Biological Networks TIAN Ye 1，2， LIU Dayou 1，2+， YANG Bo 1，2 Journal of Frontiers of Computer Science and Technology    2010, 4 (4): 330-337.   DOI: 10.3778/j.issn.1673-9418.2010.04.005 Complex networks are prevalent in the real world， they have small-world and scale-free properties. Network community structure is one of the most important topological properties of complex networks among its statistical properties. Using the clustering algorithm in complex biological networks can help us to reveal the community structure of biological networks， which is helpful to analyze the topological structures of biological networks， predict the function of community structure. This paper reviews the application and the progress of complex networks clustering algorithm used in protein-protein interaction networks and metabolic networks， analyzes the evaluation function of several clustering algorithms and their application of occasions， and discusses the major problems in clustering algorithm of the biological networks. Related Articles | Metrics

Abstract （6933）

PDF （1625）

Select

Graph Kernel Based Semi-Supervised Dimensionality Reduction Method* WU Xia+; ZHANG Daoqiang Journal of Frontiers of Computer Science and Technology    2010, 4 (7): 629-636.   DOI: 10.3778/j.issn.1673-9418.2010.07.006 Graph based data representation and analysis have received more and more attention in machine learning community. Most previous studies focus on designing an appropriate graph kernel which measures the similarity rela- tionship between graph data. Once the graph kernel is constructed, standard support vector machine (SVM) can be used for graph classification. This paper extends graph kernel methods for graph classification. It firstly performs dimensionality reduction using kernel principal component analysis (kPCA) on those high-dimensional data induced by graph kernel to obtain corresponding low-dimensional data with vector representation, and then analyzes those new data using conventional machine learning methods. Furthermore, it introduces supervision information in the form of pairwise constraints into kPCA and proposes the graph kernel based semi-supervised dimensionality reduc-tion algorithm. Experimental results on MUTAG and PTC data sets validate the effectiveness of the proposed methods. Related Articles | Metrics

Abstract （6899）

PDF （1821）

Select

Aerial Images Categorization with Deep Learning LI Xiaolong, ZHANG Zhaoxiang, WANG Yunhong, LIU Qingjie Journal of Frontiers of Computer Science and Technology    2014, 8 (3): 305-312.   DOI: 10.3778/j.issn.1673-9418.1306023 In recent decades, aerial image/video processing has been widely studied for urban planning, coastal monitoring and military tasks. Therefore, understanding the contents contained in aerial images and studying the scene classification of aerial videos are very important. However, currently most popular scene classification algorithms are mainly for natural scenes, rarely for high resolution aerial scene classification. This paper proposes a hierarchical scene classification model for aerial videos/images. Firstly, the scale-invariant feature transform (SIFT) vector is extracted as the patch feature. Then, on the basis of utilizing bag of words, the deep belief network (DBN) initialized by restricted Boltzmann machine (RBM) is used to obtain the latent variables which describe the relationship between low-level region features and high-level global features. The DBN also plays as a classifier. The proposed method achieves promising performance compared with the state of art scene classification methods. Related Articles | Metrics

Abstract （6762）

PDF （2716）

Select

A review of FEM and trend of development GU Chengzhong+, WU Xinyue Journal of Frontiers of Computer Science and Technology    2008, 2 (3): 248-259.   This paper summarizes the development of search in mesh generation for finite element computation in last ten years. Firstly， the principle of meshing is researched. Secondly， the main research fields， such as mapping methods， grid-based approach， point distribution and triangulation approach， topology decomposition approach， geometry decomposition approach， sweeping approach， are discussed and classified. Uniting instances， it analyses systematically the principle， characteristics and scopes of these methods. Thirdly， the front edge is narrated and hexahedral mesh is reviewed. Finally， this paper prospects the trend FEM. Related Articles | Metrics

Abstract （6726）

PDF （4800）