Please wait a minute...
, No 2
 ■ Cover Page (PDF 4,745 KB) ■ Editorial Board (PDF 82 KB)  ■ Table of Contents, February 2018 (98 KB)
  
  • Original articles
    Identifying Opinion Leaders with Improved Weighted LeaderRank in Online Learning Communities
    Ling Luo, You Yang, Zizhong Chen, and Yan Wei
    2018, 14(2): 193-201.  doi:10.23940/ijpe.18.02.p1.193201
    Abstract    PDF (526KB)   
    References | Related Articles
    Opinion leaders play a crucial role in closely interconnecting groups and help achieve better group performances in online learning communities. Weighted LeaderRank is superior to other methods in identifying opinion leaders, but there are some limitations in its weighted mechanism. This study further optimizes the weighted mechanism of weighted LeaderRank by introducing users’ initial comprehensive influence and the number of user interactivity. Experimental results show that the improved weighted LeaderRank algorithm can improve the accuracy of opinion leader identification in online learning communities compared with the other two typical algorithms.

    Big Data Storage and Parallel Analysis of Grid Equipment Monitoring System
    Xiaoming Zhou, Anlong Su, Guanghan Li, Weiqi Gao, Chunhua Lin, Shidong Zhu, and Zhenliu Zhou
    2018, 14(2): 202-209.  doi:10.23940/ijpe.18.02.p2.202209
    Abstract    PDF (440KB)   
    References | Related Articles

    With the analysis on data feature of grid equipment operation monitoring, this work focuses on discussing the big data storage scheme for grid equipment online monitoring data, and describes optimization measure of grid monitoring data analysis. Based on the characteristics of large data scale, multiple data types and low value density with the online monitoring data, we provide a big data storage scheme based on HDFS cloud platform using consistent hashing. Meanwhile, we also employ a multi-channel data acquisition system using multiscale multivariate entropy as the feature extraction algorithm of the multi-source power grid monitoring data. To validate the efficiency of the algorithm, we perform experiments using power grid equipment ledger data, chromatographic hydrocarbons data of transformer oil, microclimate data, and transformer vibration data for association analysis. The big data storage scheme and the feature extraction algorithm proved that it could reduce the communication overhead between storage nodes, efficiently improve system performance, and is suitable for the actual application of power grid monitoring system.

    Optimization of Software Rejuvenation Policy based on State-Control-Limit
    Weichao Dang and Jianchao Zeng
    2018, 14(2): 210-222.  doi:10.23940/ijpe.18.02.p3.210222
    Abstract    PDF (927KB)   
    References | Related Articles

    Software Rejuvenation is a proactive software control technique used to improve computing system performance when a system suffers from software aging. In this paper, a state-control-limit-based rejuvenation policy with periodical inspection has been proposed. The steady-state system availability model has been constructed based on the semi-renewal process. The steady-state probability density of degradation system stated as a function of the inspection interval and the rejuvenation threshold have been derived. The average unavailable time of when a soft failure occurs within an inspection cycle has been taken into account to calculate the steady-state system availability. Finally, the system availability with corresponding optimal inspection time interval and rejuvenation threshold have been obtained numerically.

    A Novel Image Retrieval Method with Saliency Feature Vector
    Junfeng Wu, Wenyu Qu, Zhiyang Li, and Changqing Ji
    2018, 14(2): 223-231.  doi:10.23940/ijpe.18.02.p4.223231
    Abstract    PDF (554KB)   
    References | Related Articles

    In the past few years, image retrieval has been one of the research focuses in the field of computer vision. For most retrieval methods, the accuracy of the retrieval results mainly depends on the extracted feature vectors. But, the foreground and the background in the images are not distinguished for most methods. It is obvious that these methods are not in accordance with the visual characteristics of the human eye. In this paper, salient objects are extracted from images in order to improve the pertinence of feature vector extraction. The paper utilizes a spatial pyramid model to divide the image into different parts with different scales. The feature vectors extracted in different scale are connected. Then, the saliency map and saliency score are used to rebuild the joint vector. Each feature vector is assigned different weighted values according to its different location in the image and scale. Finally, the newly constructed feature vectors are used to measure the similarity between images. In order to test the effectiveness of the algorithm, we evaluate our method on the SIMPLIcity dataset and Stanford dataset. Experimental results show that the proposed method has a great improvement in both accuracy and efficiency.

    Target Tracking based on Millimeter Wave Radar in Complex Scenes
    Guangyao Zhai, Cheng Wu, and Yiming Wang
    2018, 14(2): 232-244.  doi:10.23940/ijpe.18.02.p5.232244
    Abstract    PDF (1365KB)   
    References | Related Articles

    Currently, the method of using millimeter wave radar to detect obstacles in front of vehicles has been widely used. When using millimeter wave radar to detect obstacles on the road, the radar has more noise interference due to the changeable road environment and complex background. Combined with the complexity and variety of road targets, the random changes of scattering intensity and relative phase of different parts cause the distortion of the echo phase wave, resulting in the flicker noise that affects the accuracy of measurement, and even lead to the loss of targets. In this case, there are some shortcomings in tracking the target using the ordinary Kalman filter algorithm. In this paper, a Sage-Husa adaptive Kalman filtering algorithm is designed for the road environment to track radar targets and improve the accuracy of target tracking. Then, the radar and machine vision information fusion method is used to intuitively judge the filtering effect and determine whether the radar loses the target. Finally, the true value of the target position is approximated by filtered value when the radar loses its target. The experimental results show that this method can improve the accuracy and reliability of the millimeter wave radar.

    A Survey on Set Similarity Search and Join
    Lianyin Jia, Lulu Zhang, Guoxian Yu, Jinguo You, Jiaman Ding, and Mengjuan Li
    2018, 14(2): 245-258.  doi:10.23940/ijpe.18.02.p6.245258
    Abstract    PDF (841KB)   
    References | Related Articles

    Set similarity search and join operations have a wide range of applications, including e-commerce, information retrieval, bioinformatics, and so on. Although extensive techniques have been suggested for set similarity search and join, there is literature on systematically categorizing these techniques and comprehensively comparing them. To bridge this gap, this paper provides a comprehensive survey of set similarity search and join. The survey starts from the basic definitions, framework, ordering and similarity functions of set similarity search and join. Next, it discusses the main filtering techniques used in the state-of-the-art algorithms, and analyses the pros and cons of these algorithms in detail. For set similarity join algorithms, we divide them into 2 main categories based on the key underlying techniques they use: prefix filtering based algorithms and partition based algorithms. Prefix filtering is the most dominant technique, so algorithms based on prefix filtering and their recent variants are analyzed thoroughly. Partition is also a promising technique as it can exploit comparisons between multiple partitions. Furthermore, some efforts on MapReduce based set similarity join algorithms are also discussed briefly. In the end, the open challenges and questions are presented for future pursue.

    A Numerical Simulation of Deepwater Riser’s FSI based on MLPG
    Jianping Chen, Jie Xu, Litao Wang, Xinen Chen, and You Gong
    2018, 14(2): 259-268.  doi:10.23940/ijpe.18.02.p7.259268
    Abstract    PDF (585KB)   
    References | Related Articles

    The paper presents the numerical simulation technology of deep-water risers’ fluid-solid interaction based on mesh-free MLPG method, which applies the combination of Petrov-Galerkin Method and three-step finite element method in the mode of separating velocity and pressure to discretize control equation of two-dimensional incompressible flow field respectively in time and space. At the end of the paper, a stationary cylinder Vortex-Induced Vibration is taken as an example to be calculated and compared. It is in good agreement with the traditional numerical method based on mesh and physical model well, which proves the proposed method is effective and has good accuracy.

    Intelligent Identification of Ocean Parameters based on RBF Neural Networks
    Li Yuan, Wei Wu, Chuan Tian, Wei Song, Xinghui Cao, and Lixin Liu
    2018, 14(2): 269-279.  doi:10.23940/ijpe.18.02.p8.269279
    Abstract    PDF (2033KB)   
    References | Related Articles

    Ocean data assimilation is challenging because of interactive marine environmental parameters that are affected by macroscopic ocean dynamics. In order to overcome these challenges, a multi-variable assimilation scheme based on a Radial Basis Function (RBF) Neural Network is proposed in this paper. Relative influential parameters are considered as bounded time series variables so that they can be selected for nonlinear function approximating in the first stage. Then, a RBF Neural Network identification model is designed to simulate multiple interactive high-dimensional variables. This simulation is performed by applying proper hidden neurons. According to experimental results, this training method successfully approximates real circumstances. The identification accuracy and vibration are well constricted in the margin evaluated by 1.6×10-5.

    Relief Feature Selection and Parameter Optimization for Support Vector Machine based on Mixed Kernel Function
    Wei Zhang and Junjie Chen
    2018, 14(2): 280-289.  doi:10.23940/ijpe.18.02.p9.280289
    Abstract    PDF (392KB)   
    References | Related Articles

    In order to improve the classification performance of Support Vector Machine (SVM), Relief feature selection algorithm was used to obtain the most relevant feature subset and remove redundant features. The mixed kernel function, which combined the global kernel function with the local kernel function, was proposed to strengthen the learning ability and generalization performance of SVM. In addition, the parameter optimization of SVM, which combined Genetic Algorithm (GA) with grid search, was performed to reduce computation time and find optimal solutions. Finally, the methods presented in this paper were used in the Heart disease data set and the Breast cancer data set in the UCI. Compared with KNN and BP neural network, the classification result of SVM model with Relief algorithm and mixed kernel function significantly outperformed the other comparable classification model and the experimental results demonstrate the validity of the proposed model.

    Adaptive Short-Circuit Current Calculation Model based on Colored Petri Net
    Xinliang Wang and Xiang Jin
    2018, 14(2): 290-299.  doi:10.23940/ijpe.18.02.p10.290299
    Abstract    PDF (526KB)   
    References | Related Articles

    According to structural characteristics of the coal mine high-voltage grid, we propose an adaptive short-circuit current calculation model based on colored Petri net that includes a network topology self-learning model and a calculation model. The former model includes 10 places and 7 transitions, and the latter model includes 5 places and 4 transitions. According to coding results of the network topology self-learning model, we set the initial state of place information. Then, according to the rules of transitional trigger, the calculation model can complete adaptive short-circuit current calculation of some output switch. Simulation results show that the model can effectively learn the topology relationship of power grid and achieve adaptive short-circuit current calculation of coal mine high-voltage grid. Meanwhile, compared to the adaptive short-circuit calculation algorithm based on the state of interconnection switch, the adaptive short-circuit calculation model based on colored Petri Net can shorten the time needed to complete the short-circuit calculation. Furthermore, when the number of output switch is 72, the time overhead of short-circuit calculation is about 20 times as much as the time overhead in short-circuit calculation algorithm based on the state of interconnection switch.

    Intrusion Anomaly Detection based on Sequence
    Gangyue Lei
    2018, 14(2): 300-309.  doi:10.23940/ijpe.18.02.p11.300309
    Abstract    PDF (328KB)   
    References | Related Articles

    For single event sequences, a new anomaly detection method based on SV-LFSP (Short Variable-Length Frequent Sequence Pattern) is presented in this paper. Considering the structure character of procedure calling sequences generated by computer programs, the method defines SV-LFSP and contains three fundamental elements in the program flow, sequence, iteration and selection. To build the SV-LFSP library, the SV-LFSP generation algorithm is used. Essentially, this algorithm follows the idea of TEIRESIAS, with an additional redundancy controlling mechanism. Event flow chart, which has the capability of describing program behavior accurately, is a visual version of the SV-LFSP library. This new method is superior to previously provided frequent episode pattern matching algorithms for compact detection models, with high detection efficiency and low time delays.

    Player Detection based on Support Vector Machine in Football Videos
    Chengjun Cui
    2018, 14(2): 309-319.  doi:10.23940/ijpe.18.02.p12.309319
    Abstract    PDF (856KB)   
    References | Related Articles

    An automatic player detection method based on fuzzy decision making one-class SVM is proposed. Detection results of statistical classifier player detection methods are better than rule based player detection methods. However, manually labelled training samples are used in these statistical classifiers based player detection methods. Thus, cost is very important. To resolve this problem, we propose an instinctive player detection method using fuzzy decision making one-class SVM and automatically collected player samples. In this method, one-class SVM (OCSVM) is introduced to train the player detector by drawing lessons from the human object category classification mechanism. Additionally, decision function of OCSVM is improved by dividing the decision value dynamically using the fuzzy decision method, which is able to reduce the detection error caused by the insu?cient representativeness of the automatically collected training samples. Finally, a set of criteria is introduced to obtain the training samples automatically, and player detection experiments are performed on these training samples using FD-OCSVM. Experiments show that better detection results are obtained using the proposed method in the scenario of using automatically collected training samples, which improves the automatic degree of player detection.

    A Survey on Trajectory Big Data Processing
    Amina Belhassena and Hongzhi Wang
    2018, 14(2): 320-333.  doi:10.23940/ijpe.18.02.p13.320333
    Abstract    PDF (600KB)   
    References | Related Articles

    Rapid advancements of location-based information provided by publicly available GPS-enabled mobiles devices boost the generation of massive trajectory data. Recently, numerous researchers have addressed many problems regarding trajectory data, which is based on storage and queries processing. Further, a wide spectrum of application domains can benefit from trajectory data mining including trajectory organization as well as queries. Therefore, large-scale trajectory data has received increasing attention in research fields as well as in industry. As the massive trajectory data processing exceeds the power of centralized approaches used previously, in this paper, we survey various existing tools used to process large-scale trajectory data in a distributed way, e.g. MapReduce, Hadoop, and Spark. Furthermore, this paper reviews an extensive collection of existing applications of movement objects, including trajectory data mining and frequent trajectory. We also propose an open interesting research direction, which is challenging and has not been explored up until now, to improve the quality of trajectory query.

    A Visual Cryptography Scheme-Based DNA Microarrays
    Xuncai Zhang, Zheng Zhou, Yangyang Jiao, Ying Niu, and Yanfeng Wang
    2018, 14(2): 334-340.  doi:10.23940/ijpe.18.02.p14.334340
    Abstract    PDF (455KB)   
    References | Related Articles

    Visual cryptography is a cryptographic technique that allows visual information to be encrypted in such a way that the decryption can be performed by humans. The power of DNA molecule comes from its memory capacity and parallel processing. In this article, a visual encryption algorithm based on DNA microarrays is proposed, which successfully integrates the advantages of the algorithm in information security with the natural advantages of modern biotechnology. The algorithm converts plaintext into QR code and then uses the visual encryption scheme to encrypt the QR code image. It combines with DNA microarray technology to achieve information encryption and decryption. Security analysis shows that this algorithm has high security.

    Optimal Licensing in a Stackelberg Duopoly Market under Asymmetric Information of the Marginal Cost
    Qingyou Yan, Le Yang, Jieting Yin, and Youwei Wan
    2018, 14(2): 341-348.  doi:10.23940/ijpe.18.02.p15.341348
    Abstract    PDF (431KB)   
    References | Related Articles

    In this paper, we investigate a Stackelberg leader’s licensing behavior and its welfare consequence when the rival holds private information about the marginal cost after licensing occurs. In order to examine the effect of the asymmetric information on the optimal licensing strategy, we consider three possible forms of a two-part tariff licensing contract (excluding excluding contract, separating contract, and pooling contract). The result shows that, the optimum is either an exclusive contract with pure royalty on the low type rival or a separating contract with different royalty rates on the different type rivals, which mainly depends on the possibility that the rival is a low type. Furthermore, there is a conflict between the innovator and social welfare when the possibility that the rival is a low type is very high.

    Inferring Gender of Micro-Blog Users based on Multi-Classifiers Fusion
    Jinghua Zheng, Shize Guo, Liang Gao, Di Xue, Nan Zhao, and Huimin Ma
    2018, 14(2): 349-356.  doi:10.23940/ijpe.18.02.p16.349356
    Abstract    PDF (480KB)   
    References | Related Articles

    Knowing user demographic traits offers a great potential for public information. Most research have used local features to predict user demographic traits. Since this method did not make the most of user global features, the prediction performance was low. In this paper, our goal tries to use an ensemble learning method to improve the prediction performance through multi-classifiers fusion. Our work makes three important contributions. Firstly, we show how to predict Sina Micro-blog users’ genders based on his/her text published on the social network. Secondly, we show that user’s personality traits can also be used to infer gender. And last and thirdly, we propose multi-classifiers fusion to predict users’ genders, and give the experimental results that validate our method by comparing it with a different local features dataset. Our experiment demonstrates that our method can improve the accuracy rate, the recall rate of prediction, and the F value.

    Extracting Emotional Units based on POS Templates
    Zhenggao Pan and Lili Chen
    2018, 14(2): 357-362.  doi:10.23940/ijpe.18.02.p17.357362
    Abstract    PDF (524KB)   
    References | Related Articles

    With the increasingly popularity of electronic commerce, a large number of product reviews appeared in electronic commerce websites, which implicated a lot of valuable business information. Sentiment analysis is the core issue in disposing of business information, and the product feature words and sentiment words extraction are key technology that affect the quality of sentiment analysis. This paper proposes a simultaneous extraction algorithm of product feature words and sentiment words based on part-of-speech(POS) relation templates. Firstly, we extract possible POS dependency templates in a training set by using the supervised sequence rules mining algorithm. Secondly, we use the templates in the test samples to extract possible two tuple of product feature words and sentiment words. Finally, we test this method in a hotel review corpus. The experimental results show that this proposed method has a good application effect.

    Evaluating Impacts of Financial Risks on Schedule Delays of International Highway Projects in Vietnam using Structural Equation Model
    Hong Anh Vu, Cao Cuong Vu, Jianqiong Wang, and Lianxing Min
    2018, 14(2): 363-375.  doi:10.23940/ijpe.18.02.p18.363375
    Abstract    PDF (1056KB)   
    References | Related Articles

    Schedule delay (SD) in International Highway Projects (IHPs) has a higher risk than traditional projects and causes a significant detrimental impact on efficiency, cost, and reputation of project investment. As one of the main reasons causing SD, the financial problem was discussed in the literature but was either usually listed very generally or its impacts on project performance were insufficiently described. This study focuses on the financial-related factors affecting SD in IHPs, and employs a factor analysis technique to identify and categorize them using the 207 questionnaire data from the IHPs in Vietnam. Then, the study applies the Structural Equation Model (SEM) to assess the influences of identified factors on SD as well as interactions among them. The empirical results reveal that the financial-related risk elements of SD could be grouped into five major factors in IHPs. In accordance with the effects on SD, the five factors from highest to lowest respectively are changing economic environment(CEE), policy changes (PC), lack of financial capital (LFC), delayed payment (DP), and poor financial management (PFM). The study also discusses the implications of empirical results to effectively avoid, reduce, or even eliminate the impacts of financial-related risks on SD, and provides a set of countermeasures to help the local government and international contractors to mitigate the SD of IHPs in Vietnam, as well as in developing countries that possess similar conditions to those in Vietnam.

    Anomaly Detection based on Fuzzy Rules
    Wenjiang Jiao and Qingbin Li
    2018, 14(2): 376-385.  doi:10.23940/ijpe.18.02.p19.376385
    Abstract    PDF (422KB)   
    References | Related Articles

    Essentially, the fuzzy assert rule library is the fuzzy decision tree. A fuzzy decision tree growth algorithm based on local dynamic optimization is present. Following the idea of the greedy strategy, the algorithm ensures that once a continuous attribute is chosen as a branch node, the membership functions of this attribute after fuzzifying is dynamically optimized. On the other hand, according to fuzzy logic, enhanced Apriori algorithm is present to all the fuzzy frequent item sets composed of fuzzified attributes of multiple events. Then, the fuzzy frequent item sets are transformed into fuzzy association rules that compose the fuzzy association rule library. As for a multiple event sequence, eight different detection algorithms are provided and tested on the same platform. Experiments show that two new algorithms using the fuzzy decision tree and fuzzy association rule library detection models get the highest score.

    Target Tracking based on KCF Combining with Spatio-Temporal Context Learning
    Aili Wang, Zhennan Yang, Yushi Chen, and Yuji Iwahori
    2018, 14(2): 386-395.  doi:10.23940/ijpe.18.02.p20.386395
    Abstract    PDF (958KB)   
    References | Related Articles

    Most target tracking is based on a lot of samples training to build the model of the target, which is then carried on the tracking processing. This will need to choose a lot of tracked target samples for learning and training. However, there are all kinds of deformation of the training samples, including variety of light and scale, and so on, causing the long computation time, high computational complexity, and less robustness. The traditional kernel correlation filtering (KCF) tracking is through online learning of the first frame in the target vide. It then uses cyclic matrix to strengthen samples robustness, reducing the complexity of the calculation and time. But, the traditional KCF nuclear is unsatisfactory used for complex scenarios and quick treatment. In this paper, under the framework of the KCF, the target context information is introduced to make the tracking have better robustness and a better effect to deal with complex scenarios.

ISSN 0973-1318