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

Test clue generation based on UML interaction overview diagram ZENG Yi WANG Cuiqin LI Hanyu HONG Hao Journal of Computer Applications    2014, 34 (1): 270-275.   DOI: 10.11772/j.issn.1001-9081.2014.01.0270 Abstract （6957）      PDF （814KB）（466）       Knowledge map    Save Concerning the problem that single UML model can not test the software sufficiently, this paper proposed a new method of automatically generating software test clues by combining the characteristics of UML2.0 interaction overview diagram. First, this paper gave the formal definition of UML class diagrams, sequence diagrams and Interaction Overview Diagrams (IOD) . Second, the Node Control Flow Graph (NCFG) was constructed by extracting the process information of the interaction overview diagram while the Message Sequence Diagrams (MSD) were constructed by extracting the object interaction information. The testable model of IOD was constructed by embedding the MSD's message path into NCFG. At last, the paper adopted two-two coverage criterion to generate test clues. The experiment verifies that this method which automatically generates test clues avoids the combinatorial explosion while guaranteeing the test adequacy. Related Articles | Metrics

Select

Survey of convolutional neural network LI Yandong, HAO Zongbo, LEI Hang Journal of Computer Applications    2016, 36 (9): 2508-2515.   DOI: 10.11772/j.issn.1001-9081.2016.09.2508 Abstract （6565）      PDF （1569KB）（17814）       Knowledge map    Save In recent years, Convolutional Neural Network (CNN) has made a series of breakthrough research results in the fields of image classification, object detection, semantic segmentation and so on. The powerful ability of CNN for feature learning and classification attracts wide attention, it is of great value to review the works in this research field. A brief history and basic framework of CNN were introduced. Recent researches on CNN were thoroughly summarized and analyzed in four aspects: over-fitting problem, network structure, transfer learning and theoretic analysis. State-of-the-art CNN based methods for various applications were concluded and discussed. At last, some shortcomings of the current research on CNN were pointed out and some new insights for the future research of CNN were presented. Reference | Related Articles | Metrics

Select

Supervisal algorithm design of IDS using support vector regression Jia-chao ZHANG Journal of Computer Applications    Abstract （5365）      PDF （564KB）（1341）       Knowledge map    Save To improve classific precision of network intrusion detection model and reduce the number of training data set and learning time, a new supervisal algorithm based on ε Support Vector Regression machines (ε-SVR) machine was proposed. Firstly, normalization was used on training data set, and then a new coefficience of sparse penalty function was adjusted. Finally, the experimental results using KDD CUP 1999 data set show that this approach can detect intrusion behavior, increase its veracity and validity, and reduce its distortion. Related Articles | Metrics

Select

Research and advances on graph data mining DING Yue ZHANG Yang LI Zhan-huai WANG Yong Journal of Computer Applications    2012, 32 (01): 182-190.   DOI: 10.3724/SP.J.1087.2012.00182 Abstract （4420）      PDF （1495KB）（1512）       Knowledge map    Save With the rapid growth of bioinformatics (protein structure analysis, genome identification), social networks (links between entities), Web analysis (interlinkage structure analysis, content mining and Web log retrieval), as well as the complex structure of text information retrievals, mining graph data has become a hot research field in recent years. Some traditional data mining algorithms have been gradually extended to graph data, such as clustering, classification, and frequent pattern mining. In this paper, the authors presented several state-of-art mainstream techniques for mining graph data, and gave a comprehensive summary of their characteristics, practical significance, as well as real-life applications on mining graph data. Finally, several research directions on graph data, and particularly, uncertain graph data were pointed out. Reference | Related Articles | Metrics

Select

Cloud computing and its key techniques Journal of Computer Applications    2009, 29 (09): 2562-2567.   Abstract （4380）      PDF （931KB）（48632）       Knowledge map    Save Cloud computing is a new computing model; it is developed based on grid computing. The authors introduced the development history of cloud computing and its application situation; compared existing definitions of cloud computing and gave a new definition; took google's cloud computing techniques as an example, summed up key techniques, such as data storage technology (Google File System), data management technology (BigTable), as well as programming model and task scheduling model(Map-Reduce), used in cloud computing; and analyzed the differences among cloud computing, grid computing and traditional super-computing, and fingered out the broad development prospects of cloud computing. Related Articles | Metrics

Select

Data integration method based on SOA software architecture Journal of Computer Applications    2010, 30 (9): 2370-2373.   Abstract （4285）      PDF （649KB）（1137）       Knowledge map    Save According to the autonomy, semantic complexity and diversity of cross-platform of data, as well as combining the applications and requirements of digital business platform, a data integration method based on Service Oriented Architecture (SOA) software architecture was proposed in this paper. The proposed method, which was on the basis of Java Business Integration (JBI), could solve the problems of semantic aspect, multiple messages, and autonomy issues by means of adding four service modules. Finally, the proposed method was implemented by combining the relative techniques and system architecture of SOA. Related Articles | Metrics

Select

Application of neural network based on particle swarm algorithm for the prediction of oil quality ;; Journal of Computer Applications    Abstract （4126）      PDF （579KB）（1270）       Knowledge map    Save PSO(particle swarm optimization) algorithm is a kind of stochastic global optimization based on swarm intelligence. Through the interaction of particles, PSO searches the solution space intelligently and finds out the best solution. The advantage of PSO is that it is easy to operate and to achieve. A model integrating PSO and NN (neural network) was established in this paper, which takes full use of the global optimization of PSO and local accurate searching of BP. The example of oil quality prediction shows that PSONN is more efficient and has good generalization. Related Articles | Metrics

Select

Study on morphological model of reed and its visualization based on growth mechanism ;; Journal of Computer Applications    Abstract （3967）      PDF （859KB）（977）       Knowledge map    Save Morphogenesis model of plant considering its physiological function plays the important role in simulating plant growth on computer. After studying morphological and physiological changes of reed during their growth processes and analyzing relations between their characteristics of morphology and physiology, it introduced a method for morphological structure modeling of reed based on growth mechanism. According to this model, visualization algorithm was also presented and large numbers of data resulting from reed development were processed effectively. An example was given and experiment results illustrate the presented model is efficient in simulating reed growth. Related Articles | Metrics

Select

A Mosacing Algorithm of Large-Scale Microscope Images ;; Journal of Computer Applications    Abstract （3780）      PDF （785KB）（1052）       Knowledge map    Save Image mosaicing is an important technology in the field of image processing and computer vision. To eliminate error accumulation in the mosaicing of large-scale microscope images, a new algorithm was proposed. The algorithm solved the local alignment parameters with the hierarchical model to increase the robustness and speed of image alignment. Then it constructed the global image alignment model and gets the optimization solution by solving a set of linear equations. Experiments show that this algorithm effectively eliminates error accumulation and gets the desired result. Related Articles | Metrics

Select

Application of simulated annealing algorithm in multi-objective airline crew rostering system Journal of Computer Applications    Abstract （3678）      PDF （723KB）（1930）       Knowledge map    Save At present, the crew rostering system is based on the manual way to accomplish, and there are many restricted terms that need to be considered. Thereby it is necessary to make use of computer technology to make the crew rostering automatic and reasonable. The crew rostering system was roughly introduced, and the method about how to build models for crew rostering system was described in detail. The multi-objective of crew rostering system was achieved by simulated annealing algorithm. The actual data of airline company were used to testify its reasonableness and effectiveness. Related Articles | Metrics

Select

BPEL based method for general security control module design Guo-feng ZHANG Jun HE Cong-fu XU Journal of Computer Applications    Abstract （3567）      PDF （675KB）（1086）       Knowledge map    Save With the introduction of the Service Oriented Architecture (SOA), the integration and development speed of software systems will become quicker and quicker. But the security mechanism of software systems has to be rebuilt when they are developed. Moreover, security mechanism becomes more complex when the number of software systems increases rapidly. A Business Process Execution Language (BPEL)-based method for general security control module design was proposed, which reduced the development and management work. The operation mechanism of right module was exemplified by the account integration of Enterprise Resource Planning (ERP) in manufacturing with e-business system. Related Articles | Metrics

Select

Energy-efficient algorithm for virtual backbone construction in wireless sensor network Journal of Computer Applications    Abstract （3550）      PDF （762KB）（1646）       Knowledge map    Save An energy-efficient algorithm for virtual backbone construction that could be used for Wireless Sensor Network (WSN) routing was proposed. The network nodes were divided into a number of clusters according to the geographical distance between them, the distance of the cluster head and members was k-hop. While the size of cluster was increased, the cost of communication was reduced. Using the smallest connected dominating set theory to optimize within the cluster structure, choosing a new parameter value as a right, making the nodes with higher energy be key nodes while guarantee the size of the network backbone, the energy consumption was balanced, and the life of the network was prolonged. Simulation results show that the algorithm can reduce the network scale and routing complexity, and extend the life of WSN. Related Articles | Metrics

Select

Method for customer segmentation based on three-way decisions theory HUANG Shunliang WANG Qi Journal of Computer Applications    2014, 34 (1): 244-248.   DOI: 10.11772/j.issn.1001-9081.2014.01.0244 Abstract （3545）      PDF （752KB）（532）       Knowledge map    Save To solve the uncertainty of customer segmentation, a new method based on three-way decisions theory was proposed. The method considered the risk cost and the profit of customer segmentation comprehensively. The problem of customer segmentation was modeled based on three-way decisions theory that included computing threshold and the procedure of application. Finally, an example was given to illustrate the procedure of application and the superiority of the new method. Three-way decision method was not only used in a procedure of two-way decision, but also used independently as a decision method. In accordance with decision results of three-way decision, there were three results that can provide three different strategies for three decision domains. The introduction of three-way decision theory provides a new view for customer segmentation, which can minimize risk cost. Related Articles | Metrics

Select

An Intrusion Detection Model Based on Immune and Rough Sets Theory ;; Journal of Computer Applications    Abstract （3519）      PDF （836KB）（1035）       Knowledge map    Save Intrusion detection system has become the research hotspot because it can provide dynamic protection for computer system. Aiming at the problems existed in actual methods or models of intrusion detection, an effective method for intrusion detection based on immune theory and rough sets theory was presented in this paper. The circular sequences of system call sequences generated during the normal execution of a process is replaced by circular body, then, a little data is extracted from normal system call sequences, and is transformed to decisive table, afterward, the decisive table is reduced and the simplest rules that present normal behavior mode is extracted from reduct by rough sets theory. These rules can be used to detect anomalous behavior. In order to realize the quick detection of known intrusion, an engine of quick detection inspired by immune system theory was presented in this paper. Compared with other methods in the literature, the method presented in this paper is not only able to extract a set of effective detection rules with the minimum size from part of records of system call sequences, but also can detect the known intrusion quickly. Experiments show that this method in this paper is better than other methods based on system call. Related Articles | Metrics

Select

Linear discriminant analysis based on weighted Fisher criteria and face recognition ;; Journal of Computer Applications    Abstract （3510）      PDF （813KB）（1597）       Knowledge map    Save A novel method based on weighted discriminant analysis for face recognition was proposed in this paper. First, the Fisher criterion was redefined by introducing a weighting of the contributions of individual class pairs to the overall criterion. Then, to deal with the high dimensional and singular case in face recognition problems, a simple and efficient algorithm was developed. Finally, the proposed algorithm was tested on ORL face database, and a recognition rate of 96% was achieved by using either a common nearest neighbor classifier or a minimum distance classifier. The experimental results show our method is superior to the classical Eigenfaces and Fisherfaces. Related Articles | Metrics

Select

Design of Security Solution for Web-based Educational Administration System ;;;; Journal of Computer Applications    Abstract （3351）      PDF （823KB）（2138）       Knowledge map    Save Abstract：The security problems of traditional educational administration systems(TEASs) are analyzed. In order to improve their security, a novel integrated solution is proposed. In which, network border protection, user identification and access control, intrusion detection, servers’ security, and disaster recovery are presented or strengthened. So the invalid users can be held back, and the exceeding access of valid users can be prevented. At the same time, the new method of Identity-Based Encryption (IBE) can ensure the confidentiality, integrality, usability and the undeniable-ness of the data during the processes of storage, transmission, processing, and so on. As a result, the shortages of PKI, such as the difficulty in managing the public keys, high costs, and low performance can be overcome. Theoretical analysis and the experimental results show that it provides a good security solution to the field of MIS. Related Articles | Metrics

Select

Text zero-watermark based on use frequency of Chinese characters Journal of Computer Applications    2009, 29 (09): 2366-2368.   Abstract （3328）      PDF （626KB）（1392）       Knowledge map    Save Aiming at the problems of difficult embedding of Chinese text digital watermark, poor imperceptibleness and robustness and lack of watermark capacity, a zero-watermark algorithm based on use frequency of Chinese characters was presented. In this algorithm, the use frequency of Chinese characters in the text was calculated, and then combining the frequency table of commonly used Chinese characters, the text characteristic was extracted to construct a zero-watermark. According to the trait of the watermark algorithm, a correlation function was defined to fix threshold value and examine the watermark. Experimental results indicate that the watermark algorithm constructed in this way has good imperceptibleness, robustness and sufficient watermark capacity, and the constructed watermark examining algorithm based on the correlation function exhibits comparatively lower misjudging rate. Related Articles | Metrics

Select

LU matrix key pre-distribution scheme with authentication Chow Andy weidong qiu mi wen Journal of Computer Applications    Abstract （3305）      PDF （552KB）（1459）       Knowledge map    Save The key pre-distribution based on LU composition of symmetric matrix is an effective approach to the Key management in the Wireless Sensor Networks (WSN). But there exists the possibility of failing to build the common communication key. To fix this defect and decrease communication traffic, a key pre-distribution scheme with authentication mechanism was proposed, which was based on the LU matrix scheme. Through optimizing the key pre-distribution scheme, we prove that the average communication traffic decreases by 16.7%. At last this scheme’s implementation in the Sun Spot Sensor and a detailed performance analysis were given out. Related Articles | Metrics

Select

Knowledge graph survey： representation， construction， reasoning and knowledge hypergraph theory TIAN Ling, ZHANG Jinchuan, ZHANG Jinhao, ZHOU Wangtao, ZHOU Xue Journal of Computer Applications    2021, 41 (8): 2161-2186.   DOI: 10.11772/j.issn.1001-9081.2021040662 Abstract （3261）      PDF （2811KB）（4259）       Knowledge map    Save Knowledge Graph (KG) strongly support the research of knowledge-driven artificial intelligence. Aiming at this fact, the existing technologies of knowledge graph and knowledge hypergraph were analyzed and summarized. At first, from the definition and development history of knowledge graph, the classification and architecture of knowledge graph were introduced. Second, the existing knowledge representation and storage methods were explained. Then, based on the construction process of knowledge graph, several knowledge graph construction techniques were analyzed. Specifically, aiming at the knowledge reasoning, an important part of knowledge graph, three typical knowledge reasoning approaches were analyzed, which are logic rule-based, embedding representation-based, and neural network-based. Furthermore, the research progress of knowledge hypergraph was introduced along with heterogeneous hypergraph. To effectively present and extract hyper-relational characteristics and realize the modeling of hyper-relation data as well as the fast knowledge reasoning, a three-layer architecture of knowledge hypergraph was proposed. Finally, the typical application scenarios of knowledge graph and knowledge hypergraph were summed up, and the future researches were prospected. Reference | Related Articles | Metrics

Select

An Immune Genetic Algorithm Based on Chaos Theory Nan Zhang;;; Journal of Computer Applications    Abstract （3165）      PDF （521KB）（1045）       Knowledge map    Save To increase operating efficiency of immune genetic algorithm, a chaos immune genetic algorithm is presented based on the chaotic character of immune system. Over-spread character of chaos sequence was used to overcome redundancies, and chaos initial value sensitivity was used to enlarge search space. Thus, the diversity of population was retained and local optimization was avoided. The experimental results show that the algorithm evidently improves the convergent speed and astringency. Related Articles | Metrics

Select

Improved approach to CHI in feature extraction Journal of Computer Applications    Abstract （3145）      PDF （461KB）（1447）       Knowledge map    Save Feature extraction technology is an essential part of text categorization, which directly affects the categorization precision. This paper comprehensively took frequency, distribution and concentration into account and proposed an improved Chisquare Statistic(CHI) approach. In order to verify the improved CHI approach, a contrastive experiment was carried out. The experimental results show that improved CHI approach is superior to traditional CHI approach in feature selection, which verifies the efficiency and probability of the improved CHI approach. Related Articles | Metrics

Select

Foundation excavation co-evolution based on particle swarm optimization Chen Qiu-lian Tao-shen LI Journal of Computer Applications    Abstract （3139）      PDF （806KB）（806）       Knowledge map    Save As a multilevel optimization problem, foundation excavation optimization concerns many constraint conditions. With the analysis in hierarchy and relation of level design, a new collaborative optimization algorithm was proposed, which was distributed, hierarchical and of coevolution. Collaborative optimization was integrated with particle swarm optimization to speed up the evolution and simply the optimization process. Collision and disposal in optimization of subsystems, which belong to coevolution optimization system, were analyzed. At last, an experiment of parallel collaborative design with "anchored piles in row and waterproof curtain" was given to show its effectiveness. Related Articles | Metrics

Select

Text sentiment analysis based on feature fusion of convolution neural network and bidirectional long short-term memory network LI Yang, DONG Hongbin Journal of Computer Applications    2018, 38 (11): 3075-3080.   DOI: 10.11772/j.issn.1001-9081.2018041289 Abstract （3088）      PDF （906KB）（1849）       Knowledge map    Save Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN) are widely used in natural language processing, but the natural language has a certain dependence on the structure, only relying on CNN for text classification will ignore the contextual meaning of words, and there is a problem of gradient disappearance or gradient dispersion in the traditional RNN, which limits the accuracy of text classification. A feature fusion model for CNN and Bidirectional Long Short-Term Memory (BiLSTM) was presented. Local features of text were extracted by CNN and global features related to text were extracted by BiLSTM network. The features extracted by the two complementary models were merged to solve the problem of ignoring the contextual semantic and grammatical information of words in a single CNN model, and the fusion model also effectively avoided the problem of gradient disappearance or gradient dispersion in traditional RNN. The experimental results on two kinds of datasets show that the proposed fusion feature model can effectively improve the accuracy of text classification. Reference | Related Articles | Metrics

Select

Review of event causality extraction based on deep learning WANG Zhujun, WANG Shi, LI Xueqing, ZHU Junwu Journal of Computer Applications    2021, 41 (5): 1247-1255.   DOI: 10.11772/j.issn.1001-9081.2020071080 Abstract （3082）      PDF （1460KB）（4007）       Knowledge map    Save Causality extraction is a kind of relation extraction task in Natural Language Processing (NLP), which mines event pairs with causality from text by constructing event graph, and play important role in applications of finance, security, biology and other fields. Firstly, the concepts such as event extraction and causality were introduced, and the evolution of mainstream methods and the common datasets of causality extraction were described. Then, the current mainstream causality extraction models were listed. Based on the detailed analysis of pipeline based models and joint extraction models, the advantages and disadvantages of various methods and models were compared. Furthermore, the experimental performance and related experimental data of the models were summarized and analyzed. Finally, the research difficulties and future key research directions of causality extraction were given. Reference | Related Articles | Metrics

Select

Efficient fractal image compression based on pixels distribution and triangular segmentation Journal of Computer Applications    2010, 30 (2): 337-340.   Abstract （3081）      PDF （712KB）（1584）       Knowledge map    Save The baseline fractal image compression algorithm requires a great deal of time to complete the encoding process. In order to solve this problem, an efficient fractal image compression based on pixels distribution and triangular segmentation was proposed in this paper. Exploiting the characteristics of centroid uniqueness and centriod position invariance in discrete particulate system, the matching between range blocks and domain blocks was implemented. In addition, dividing original image into isosceles right triangles segmentation could reduce the number of domain blocks, and raise the efficiency of fractal coding. Experimental results show that the proposed algorithm can produce reconstructed images with good quality and require one third time of the baseline encoding algorithm. Related Articles | Metrics

Select

Research and Implementation on Routing Mechanism Based on Interest Mining in unstructured P2P Systems YiHong Tan;; Journal of Computer Applications    Abstract （3067）      PDF （548KB）（1117）       Knowledge map    Save In the environment of unstructured peer-to-peer (P2P), routing scheme is one of the key factors affecting information searches. A routing scheme based on interest indexical table is proposed in this paper. Moreover, on the basis of it, we implement a P2P full text information retrieval prototype system Isearch. Firstly, we represent local file content of a peer with vector space model. After that, vector space is clustered to obtain interest class of this peer. And then, interest similar degree is computed among peers to build interest indexical table locally. When search occurs, query requests are forwarded to the peers with similar interest directly according to the interest indexical table. Experimental results show that ISearch can not only make good retrieval results, but also reduce the number of query peers and make retrieval more efficient. Related Articles | Metrics

Select

Deep packet inspection algorithm based on regular expressions Journal of Computer Applications    Abstract （3056）      PDF （780KB）（21395）       Knowledge map    Save This paper proposed a new DFA-based pattern matching algorithm. Based on the analysis of the impact of the number of DFA states on the algorithm performance, further improvement to the algorithm was made by introducing a DFA state number optimization algorithm. The proposed algorithm has been implemented in Linux environment and lots of experiments have been done. Experimental results show that the performance of the proposed algorithm is much better than others. Related Articles | Metrics

Select

Artificial bee colony algorithm based on chaos local search operator Wang Xiang LI Zhi-yong XU Guo-yi WANG Yan Journal of Computer Applications    2012, 32 (04): 1033-1036.   DOI: 10.3724/SP.J.1087.2012.01033 Abstract （3037）      PDF （730KB）（618）       Knowledge map    Save In order to improve the ability of Artificial Bee Colony (ABC) algorithm at exploitation, a new Chaos Artificial Bee Colony (CH-ABC) algorithm was proposed for continuous function optimization problems. A new chaotic local search operator was embedded in the framework of the new algorithm. The new operator, whose search radius shrinks with the evolution generation, can do the local search around the best food source. The simulation results show that: compared with those of ABC algorithm, the solution quality and the convergence speed of the new algorithm are better for Rosenbrock and the convergence speed of the new algorithm is better for Griewank and Rastrigin. Reference | Related Articles | Metrics

Select

Federated learning survey：concepts， technologies， applications and challenges Tiankai LIANG, Bi ZENG, Guang CHEN Journal of Computer Applications    2022, 42 (12): 3651-3662.   DOI: 10.11772/j.issn.1001-9081.2021101821 Abstract （2994）   HTML （205）    PDF （2464KB）（2213）       Knowledge map    Save Under the background of emphasizing data right confirmation and privacy protection， federated learning， as a new machine learning paradigm， can solve the problem of data island and privacy protection without exposing the data of all participants. Since the modeling methods based on federated learning have become mainstream and achieved good effects at present， it is significant to summarize and analyze the concepts， technologies， applications and challenges of federated learning. Firstly， the development process of machine learning and the inevitability of the appearance of federated learning were elaborated， and the definition and classification of federated learning were given. Secondly， three federated learning methods （including horizontal federated learning， vertical federated learning and federated transfer learning） which were recognized by the industry currently were introduced and analyzed. Thirdly， concerning the privacy protection issue of federated learning， the existing common privacy protection technologies were generalized and summarized. In addition， the recent mainstream open-source frameworks were introduced and compared， and the application scenarios of federated learning were given at the same time. Finally， the challenges and future research directions of federated learning were prospected. Table and Figures | Reference | Related Articles | Metrics

Select

Locally adaptive image denoising based on bivariate shrinkage function Xin Liu;; Journal of Computer Applications    Abstract （2964）      PDF （808KB）（1094）       Knowledge map    Save To improve the performance of image-denoising methods, a locally adaptive denoising algorithm was presented. The new algorithm assumed the statistical dependence among wavelet coefficients. First, a bivariate probability distribution model was introduced to model the statistics of wavelet coefficients, and corresponding nonlinear threshold function (bivariate shrinkage function) was derived from the model using the Bayesian estimation theory. Secondly, using locally variance estimation, a locally adaptive image-denoising algorithm was presented. Also this algorithm could be applied to the complex wavelet domain. Experimental results and comparision analysis are given to illustrate the effectiveness of this denoising algorithm. Related Articles | Metrics

Select

New reduction strategy of large-scale training sample set for SVM Journal of Computer Applications    2009, 29 (10): 2736-2740.   Abstract （2943）      PDF （950KB）（1665）       Knowledge map    Save It has become a bottleneck to use Support Vector Machine (SVM) due to such problems as slow learning speed, large buffer memory requirement, low generalization performance and so on, which are caused by large-scale training sample set and outlier data immixed in the other class. Concerning these problems, this paper proposed a new reduction strategy for large-scale training sample set according to the analysis on the structure of the training sample set based on the point set theory. This new strategy gets the potential support vectors and removes the non-boundary outlier data immixed in the other class by using fuzzy clustering. That can greatly reduce the scale of the training sample set and improve the generalization performance by effectively avoiding over-learning caused by outlier data, and finally speed up learning rate without reducing the classification accuracy. Related Articles | Metrics

Select

Cryptanalysis and improvement of a proxy signature scheme Journal of Computer Applications    Abstract （2936）      PDF （465KB）（886）       Knowledge map    Save The security of the Fu-Kou-Xiao’s proxy signature scheme with proxy signer’s privacy anonymity was analyzed, and it was shown that the scheme did not possess the property of strong unforgeability. A forgery attack was given. Using this attack, a dishonest original signer can forge a proxy signing key and produce valid proxy signatures. The reason why the attack can work was analyzed and an improved scheme was proposed to remove the attack. Related Articles | Metrics

Select

Hierarchical categorization methods of Chinese text based on vector space model ; Journal of Computer Applications    Abstract （2933）      PDF （641KB）（1263）       Knowledge map    Save On large amount conditions of text quantity, hierarchical text categorization was an effective approach. Aiming at structural characteristics of hierarchical text categorization, and considering various demands of texts in different levels on both feature selection and categorization method, a new method, Feature Dual-Selection(FDS), and an algorithm of Hierarchical Text Categorization(HTC) based on vector space model was proposed. FDS is to perform feature selection in each level, and then modify feature number along with term weighting method accordingly; HTC algorithm integrates together center classification method and Support Vector Machine(SVM), which proves more effective for broad classification and subdivision respectively. Finally, experiment results show that the new approach, proposed in this paper, outperforms plain or generic hierarchical methods with improved accuracy. Related Articles | Metrics

Select

Overview on wireless mesh networking and mobile transition techniques in IEEE 802 series FANG Xu-ming Cai-Xia QI Zheng XIANG Journal of Computer Applications    Abstract （2909）      PDF （1273KB）（1342）       Knowledge map    Save When constructing large scale wireless area networks, the traditional network techniques become unaccommodated due to lack of flexible extension and fast transition support. The IEEE standard organizations are devoted to advancing the technology of mesh networking and fast transition or mobility support. The mesh networking and transition or mobility support techniques were discussed based on Wireless Local Area Network(WLAN), Wireless Personal Area Network(WPAN) and Wireless Municipal Area Network(WMAN). The development tendency of wireless area networks was introduced to benefit the research and application of the wireless multihop networks. Related Articles | Metrics

Select

A New Approach of Image Edge Detection Based on Sobel Operators and Grey Relation KangTai Wang; Journal of Computer Applications    Abstract （2903）      PDF （886KB）（2071）       Knowledge map    Save A new approach to detect image edge using the Sobel operators and grey relation was brought forward, and then mechanism and algorithm were introduced. Simulation shows that this method has a high precision in image detection, a better antinoise ability and improving image detection effect. Related Articles | Metrics

Select

Algorithm for choosing ARQ feedback types based on IEEE 802.16 CAI Cang-Fu Yan YANG Journal of Computer Applications    Abstract （2903）      PDF （583KB）（680）       Knowledge map    Save ARQ mechanism in IEEE 802.16 can resolve the problems for data transmission on wireless link, but there is no appropriate solution for choosing the ARQ feedback types. The key to the algorithm is to choose an appropriate feedback type to send feedback messages for high resource utilization according to wireless link. Therefore, a new algorithm for choosing proper feedback types in real time was proposed. The simulation results show that the new algorithm can increase resource utilization. Related Articles | Metrics

Select

Task scheduling algorithm based on improved genetic algorithm in cloud computing environment Journal of Computer Applications    2011, 31 (01): 184-186.   Abstract （2870）      PDF （435KB）（86099）       Knowledge map    Save The number of user is huge in Cloud Computing, and the number of tasks and the amount of data are also huge. How to schedule tasks efficiently is an important issue to be resolved in Cloud Computing environment. An Double-Fitness Genetic Algorithm (DFGA) is brought up for the programming framework of Cloud Computing. Through this algorithm the better task scheduling result which has not only shorter total-task-completion time and also has shorter average-completion time can be found out. There is a contrast between DFGA and Adaptive Genetic Algorithm (AGA) through simulation experiment, and the result is: the DFGA is better, it is an efficiently task scheduling algorithm in Cloud Computing environment. Related Articles | Metrics

Select

Neighborhood aware routing protocol based on AODV ZhongCan Zhao; Journal of Computer Applications    Abstract （2856）      PDF （382KB）（1036）       Knowledge map    Save Every node has its neighbors in ad hoc networks. Making use of the information of neighbors, the routing protocol can be optimized observably. In this paper, a new Ad Hoc routing protocol named NA-AODV based on AODV is proposed. Taking advantage of neighbors, NA-AODV enable the routing nodes to repair the broken route quickly. The simulation analysis indicates that the throughput and end-to-end delay of NA-AODV are much better than that of AODV. Related Articles | Metrics

Select

Attention mechanism based pedestrian trajectory prediction generation model SUN Yasheng, JIANG Qi, HU Jie, QI Jin, PENG Yinghong Journal of Computer Applications    2019, 39 (3): 668-674.   DOI: 10.11772/j.issn.1001-9081.2018081645 Abstract （2844）      PDF （1160KB）（1476）       Knowledge map    Save Aiming at that Long Short Term Memory (LSTM) has only one pedestrian considered in isolation and cannot realize prediction with various possibilities, an attention mechanism based generative model for pedestrian trajectory prediction called AttenGAN was proposed to construct pedestrian interaction model and predict multiple reasonable possibilities. The proposed model was composed of a generator and a discriminator. The generator predicted multiple possible future trajectories according to pedestrian's past trajectory probability while the discriminator determined whether the trajectories were really existed or generated by the discriminator and gave feedback to the generator, making predicted trajectories obtained conform social norm more. The generator consisted of an encoder and a decoder. With other pedestrians information obtained by the attention mechanism as input, the encoder encoded the trajectories of the pedestrian as an implicit state. Combined with Gaussian noise, the implicit state of LSTM in the encoder was used to initialize the implicit state of LSTM in the decoder and the decoder decoded it into future trajectory prediction. The experiments on ETH and UCY datasets show that AttenGAN can provide multiple reasonable trajectory predictions and can predict the trajectory with higher accuracy compared with Linear, LSTM, S-LSTM (Social LSTM) and S-GAN (Social Generative Adversarial Network) models, especially in scenes of dense pedestrian interaction. Visualization of predicted trajectories obtained by the generator indicated the ability of this model to capture the interaction pattern of pedestrians and jointly predict multiple reasonable possibilities. Reference | Related Articles | Metrics

Select

Method of weather recognition based on decision-tree-based SVM Li Qian FAN Yin ZHANG Jing LI BAOqiang Journal of Computer Applications    2011, 31 (06): 1624-1627.   DOI: 10.3724/SP.J.1087.2011.01624 Abstract （2802）      PDF （620KB）（872）       Knowledge map    Save To improve the quality of video surveillance outdoors and to automatically acquire the weather situations, a method to recognize weather situations in outdoor images is presented. It extracted such parameters as power spectrum slope, contrast, noise, saturation as features to realize the multi-classification of weather situations with Support Vector Machine (SVM). Then a decision tree was constructed in accordance with the distance between these features. The experimental results on WILD image base and our image set of eight hundred samples show that the proposed method can recognize sunny, overcast, foggy weather more than 85%, and recognize rainy weather more than 75%. Related Articles | Metrics