Loading...

Table of Content

      
    Machine Learning & Data Mining
    The diagnosis of Alzheimer's disease classification based on multi-scale residual neutral network
    Zhenbing LIU,Xusheng FANG,Huihua YANG,Rushi LAN
    Journal of Shandong University(Engineering Science). 2018, 48(6):  1-7, 18.  doi:10.6040/j.issn.1672-3961.0.2018.205
    Abstract ( 764 )   HTML ( 143 )   PDF (2154KB) ( 438 )   Save
    Figures and Tables | References | Related Articles | Metrics

    A multi-scale resnet (MSResnet) method was proposed in this paper, which employed multi-scale convolution kernel to extract multi-scale information of structural magnetic resonance imaging MRI, and carried out residual learning for neural network, so as to avoid network degradation. After the gray scale standardization of MRI, the 99.41% classification precision was obtained by using the MSResnet model between Alzheimer's disease (AD) and normal control (NC), and the classification accuracy between AD and mild cognitive impairment (MCI) was 97.35%. Compared with the existing approaches, the algorithm proposed in this paper improved the classification accuracy significantly.

    A short text dynamic clustering approach bias on new topic
    Yingxue ZHU,Ruizhang HUANG,Can MA
    Journal of Shandong University(Engineering Science). 2018, 48(6):  8-18.  doi:10.6040/j.issn.1672-3961.0.2018.193
    Abstract ( 696 )   HTML ( 11 )   PDF (3212KB) ( 275 )   Save
    Figures and Tables | References | Related Articles | Metrics

    The dynamic Dirichlet multinomial mixture (DDMM) model for short textual data stream dynamic clustering problem was proposed.The model could capture the change of topics in the short textual data stream over time, and take the relationship between existing historical topics and new topics into consideration, which could adjust the strength of the lineage of topics, and increase the likelihood of new topic emergence.In addition, the proposed approach could infer the number of clusters automatically in the process of Gibbs sampling.Experiments indicated that the DDMM model performed well on the synthetic data set as well as real data sets.And the comparison between the proposed approach and state-of-the-art dynamic clustering approaches showed that the DDMM model was effective for document dynamic clustering, and performed well on short text dynamic clustering.

    Genetic algorithm based on Grefenstette coding and 2-opt optimized
    Xiaoyan GONGYE,Peiguang LIN,Weilong REN
    Journal of Shandong University(Engineering Science). 2018, 48(6):  19-26.  doi:10.6040/j.issn.1672-3961.0.2018.203
    Abstract ( 625 )   HTML ( 10 )   PDF (1472KB) ( 220 )   Save
    Figures and Tables | References | Related Articles | Metrics

    Grefenstette coding and 2-opt were applied simultaneously into genetic algorithm to obtain the space searching path, using a certain number of city coordinates. The synthetic experimentation achieved good result: the optimal path could be approximately represented by initial path, and could be accurately achieved via 122 iterations. And this result demonstrated that the proposed solution of search space path enabled genetic algorithm quickly converge to the optimal solution, maintained a strong search capability, achieved global optimization, and prevented local optimum.

    Multi-label feature selection algorithm based on correntropy andmanifold learning
    Hong CHEN,Xiaofei YANG,Qing WAN,Yingcang MA
    Journal of Shandong University(Engineering Science). 2018, 48(6):  27-36.  doi:10.6040/j.issn.1672-3961.0.2018.264
    Abstract ( 516 )   HTML ( 5 )   PDF (2236KB) ( 199 )   Save
    Figures and Tables | References | Related Articles | Metrics

    A sparse regularization method based on correntropy and feature manifold learning was proposed to solve the problem of multi-label feature selection. A regression model of multi-label feature selection was presented by means of correntropy. The sparse regularized multi-label feature selection model, combing ?2, 1 norm and feature manifold learning, was established. An iterative algorithm was proposed for the above model. The convergence of the algorithm was proved and the effectiveness of the given algorithm was verified through experiments.

    A parallel adaptive news topic tracking algorithm based on N-Gram language model
    Qingtao QU,Qicheng LIU,Chunxiao MU
    Journal of Shandong University(Engineering Science). 2018, 48(6):  37-43.  doi:10.6040/j.issn.1672-3961.0.2018.204
    Abstract ( 789 )   HTML ( 16 )   PDF (1102KB) ( 184 )   Save
    Figures and Tables | References | Related Articles | Metrics

    When the traditional vector space model and unigram model expressed the text features of the topic, the word order relations between the words was ignored. In terms of this issue, a parallel adaptive news topic tracking algorithm based on N-Gram language model was proposed. N-Gram language mode was used to express the text features, which made use of word order relations in news reports. The Bayes classification algorithm was applied to conduct topic tracking, with the minimum feature average confidence threshold update strategy, the training set was updated to improve the topic model by using the test news reports. The parallel adaptive news topic tracking algorithm based on N-Gram language model (PATT-Gram) was implemented on the mapreduce distributed computing model. Experiments showed that the algorithm effectively improved the topic tracking effect and had good parallel speedup and scalability.

    An adaptive ensemble classification method based on deep attribute weighting for data stream
    Yao LI,Zhihai WANG,Yan′ge SUN,Wei ZHANG
    Journal of Shandong University(Engineering Science). 2018, 48(6):  44-55, 66.  doi:10.6040/j.issn.1672-3961.0.2018.198
    Abstract ( 626 )   HTML ( 9 )   PDF (1813KB) ( 618 )   Save
    Figures and Tables | References | Related Articles | Metrics

    Due to most of the existing data stream ensemble classification algorithms without considering the importance of historical data in the evaluation of the base classifier, while ignoring the treatment of interference with irrelevant attributes and noise attributes, an adaptive ensemble classification method based on deep attribute weighting for data stream (EMDAW) was proposed to effectively combine multiple naive Bayesian models based on depth attribute weighting. In different data blocks, the contribution of different attribute values to the attribution of class attributes was deeply analyzed, and the learned local attribute weights to different attribute values were applied to reduce noise data interference. In the evaluation of the base classifier, the importance of the historical data and the current latest data was weighed. The sub-classifier combination was used to improve the overall classification performance by using the combined voting strategy based on the test case classifier confidence and classification correct rate. By comparing experiments with classical algorithms on multiple benchmark datasets, the proposed algorithm had certain advantages in classification correct rate and concept drift adaptability.

    Multi-node human behavior recognition based on linear acceleration
    Xing LI,Zhenjie HOU,Jiuzhen LIANG,Xingzhi CHANG
    Journal of Shandong University(Engineering Science). 2018, 48(6):  56-66.  doi:10.6040/j.issn.1672-3961.0.2018.202
    Abstract ( 569 )   HTML ( 6 )   PDF (4138KB) ( 230 )   Save
    Figures and Tables | References | Related Articles | Metrics

    Focused on the issue that the behavior data in the current acceleration-based human behavior recognition method was affected by the gravitational acceleration and the lack of spatial information, a multi-node human behavior recognition algorithm based on linear acceleration was proposed. The linear acceleration was obtained by removing gravitational acceleration using segmented bi-directionally removal of gravitational acceleration algorithm. The tremor motion signal was filtered by a sliding averaging filter for linear acceleration and sensor acceleration, and the redundant actions in the two accelerations were cropped. The dynamic time warping (DTW) distance features between different joint points and seven conventional time domain features were extracted from two accelerations. The support vector machine was employed to recognize the human behavior. Experimental results showed that this method could effectively improve the accuracy of human behavior recognition.

    Person re-identification based on random erasing pedestrian alignmentnetwork method
    Cui JIN,Hongyuan WANG,Shoubing CHEN
    Journal of Shandong University(Engineering Science). 2018, 48(6):  67-73.  doi:10.6040/j.issn.1672-3961.0.2018.192
    Abstract ( 908 )   HTML ( 10 )   PDF (4347KB) ( 395 )   Save
    Figures and Tables | References | Related Articles | Metrics

    The detected pedestrian images were prone to misalignment and the depth network was prone to over-fitting phenomenon. Pedestrian datasets were preprocessed using pedestrian alignment networks and random erasing data enhancements. It made the images generating different levels of occlusion, and corrected the misalignment in the images by the spatial transformation network layer in the affine estimation branch. It cropped the large part of the background and filled in the missing part of the pedestrian images, which reduced the phenomenon of network over-fitting and improved the generalization ability of the network. The tests were performed on the Market1501, DuckMTMC-reID and CUHK03 datasets, which showed the value of rank-1 reached approximately 84%. Compared the methods of randomly erasing pedestrian alignment network with other methods, it was found that the test results of pedestrian recognition method for randomly erasing pedestrian alignment network were better.

    Linguistic concept formal decision context analysis based on granular computing
    Kuo PANG,Siqi CHEN,Xiaoying SONG,Li ZOU
    Journal of Shandong University(Engineering Science). 2018, 48(6):  74-81.  doi:10.6040/j.issn.1672-3961.0.2018.208
    Abstract ( 766 )   HTML ( 7 )   PDF (1036KB) ( 205 )   Save
    Figures and Tables | References | Related Articles | Metrics

    Aiming at the decision problem with linguistic value information, combining linguistic terminology and information system, the linguistic decision information system and linguistic concept was proposed, and the related properties of linguistic concept were discussed. By transforming linguistic decision information systems, the linguistic concept formal decision context was proposed. In order to expand the intent and extent of the linguistic concept lattice, the scale of the linguistic concept was compressed, and the granular linguistic concept formal decision context was proposed by granulating the linguistic concept. The granular computing was introduced into the granular linguistic concept formal decision context, and a rule extraction model based on granular computing for linguistic concept formal decision context was constructed by using coverage and confidence. Medical diagnostic examples illustrated the effectiveness and utility of this method in obtaining high quality rules.

    An improved Apriori algorithm based on compression matrix
    Fang LIU,Guangchao WU
    Journal of Shandong University(Engineering Science). 2018, 48(6):  82-88.  doi:10.6040/j.issn.1672-3961.0.2018.206
    Abstract ( 740 )   HTML ( 11 )   PDF (2097KB) ( 228 )   Save
    Figures and Tables | References | Related Articles | Metrics

    An improved Apriori algorithm was proposed to solve the problem that the tradditional Apriori algorithm needed scan transaction database frequently and generate a large number of candidate item sets. This algorithm adopted the idea of matrix compression and added three vectors that were respectively used to represent the number of 1 in rows and columns in the transaction matrix, namely, number of transaction items and number of project support, and number of repeated transaction occurrence, so as to reduce the matrix size and avoid scanning the database multiple times. In the process of matrix operation, the number of transaction items and the number of project support in the matrix were sorted and the unsatisfied item sets and infrequent item sets were deleted to form a new matrix structure and improve the spatial efficiency of space. Performance analysis and experimental analysis of the improved algorithm showed that the algorithm was more efficient than Apriori algorithm and could mine frequent item sets more effectively.

    Segmentation of connected characters based on improved drop-fall algorithm
    Qiyue SONG,Xuewen MU,Huan CHENG
    Journal of Shandong University(Engineering Science). 2018, 48(6):  89-94, 108.  doi:10.6040/j.issn.1672-3961.0.2018.199
    Abstract ( 1693 )   HTML ( 76 )   PDF (3250KB) ( 630 )   Save
    Figures and Tables | References | Related Articles | Metrics

    As the traditional segmentation methods could not segment connected characters correctly, a segmentation algorithm based on improved drop-fall algorithm was proposed. The algorithm included two steps. Zhang-Sueng's thinning algorithm and the clustering of the connected region via self-organizing maps was used to find the starting drop point of drop-fall algorithm. A new drop path was defined to improve drop-fall algorithm. The water dropped from the starting drop point, along the skeleton of the character overlap stroke, at the end of the overlapped stroke skeleton, then continued dropping along the slant angle direction of the skeleton, until met the boundary of the character connected part. The water drop path was defined as the connected character segmentation path. This method solved the problem of character strokes fracture caused by the traditional drop-fall algorithm. Compared with the traditional drop-fall algorithm and the vertical projection segmentation algorithm, the experimental results showed that it was an ideal method for segmenting connected characters.

    Electrical Engineering
    Novel traveling wave fault location method using overdetermined equations
    Youmin LI,Shouguo LÜ,Yang ZHOU,Yaxin NIE,Mingliang JIA,Bin BI
    Journal of Shandong University(Engineering Science). 2018, 48(6):  95-100.  doi:10.6040/j.issn.1672-3961.0.2018.267
    Abstract ( 687 )   HTML ( 8 )   PDF (1283KB) ( 218 )   Save
    Figures and Tables | References | Related Articles | Metrics

    The uncertainty of traveling wave velocity brought the fault location measure error. In order to eliinate the influence of wave velocity, a two-terminal traveling wave fault location method for transmission lines based on overdetermined equation was proposed. This method first needed measure the arriving time of three waves, which were the travelling wave from the fault location to two terminals of the transmission lines, and the two refection waves between the opposite bus and the measure bus. Then with the parameters of the time and lines length, the overdetermined equations were constructed, whose solution could eliminate the uncertainty of traveling wave velocity. The simulation results showed that this method could effectively eliminate the influence of wave velocity, fault distance and ground resistance, and meanwhile show higher location precision and reliability in comparison with other traditional methods.

    Wind and PV installed capacity optimization with hybrid uncertainty of renewable energy
    Shizhan SONG,Chuanyong WANG,Wenwen KANG,Jian ZHANG,Honghua YAN,Peng LI
    Journal of Shandong University(Engineering Science). 2018, 48(6):  101-108.  doi:10.6040/j.issn.1672-3961.0.2018.289
    Abstract ( 576 )   HTML ( 10 )   PDF (3864KB) ( 231 )   Save
    Figures and Tables | References | Related Articles | Metrics

    By mining the daily and monthly characteristics of renewable energy, a time series model of wind and PV was presented. Through the Monte Carlo method, several wind and PV time series were produced to depict the hybrid uncertainty of the wind and PV. Based on the hybrid uncertainty constraints of wind and PV, power balance constraint and allowed wind and PV curtailment rate, etc., the profit maximum model of renewable was proposed to optimize the installed capacities of wind power and PV, to ensure the operation economics of wind power and photovoltaic system. The simulations showed that under different allowed wind and PV curtailment rates, the optimal installation capacities of wind power and PV varied significantly. Due to the unique daily characteristics and high installation cost of PV system, the higher total revenue of PV could be achieved with lower curtailment rate and installation capacity of PV systems.

    Restrictions and countermeasures of distributed PV integration based on secondary battery energy storage system
    Fei WANG,Chunyi WANG,Chuanyong WANG,Guangfeng ZHAO,Mu LI,Xiaohan SHI
    Journal of Shandong University(Engineering Science). 2018, 48(6):  109-115, 121.  doi:10.6040/j.issn.1672-3961.0.2018.260
    Abstract ( 547 )   HTML ( 6 )   PDF (6513KB) ( 363 )   Save
    Figures and Tables | References | Related Articles | Metrics

    The distributed PV in rural areas was taken as object, the main factors limiting the integration of the PV generation was analyzed and the feasible solution was proposed based on secondary battery energy storage system. The weakness of distributed PV integration was analyzed from the points of view of primary equipment capacity, voltage control and fluctuation of PV output power. The investment model of grid upgrading and energy storage installation was built, based on which the economic comparison was carried out and the countermeasure based on secondary battery energy storage system was proposed. A prototype of secondary battery energy storage system adopting dynamic group theory was developed and the field test was carried out. The results demonstrated that the prototype could reduce the variance of the photovoltaic power by 35% as well as reduced the maximum peak and peak difference by 60%. The prototype could also reduce the peak and valley difference of the transformer by 20% and the load rate by 12%, which could effectively support the distribution of distributed PV.

    Others
    Influence of internal pressure on the safety of exhaust duct under multiple loads
    Longlin WANG,Xundong HU,Jinsheng QI,Chunguo AN,Zhan WANG
    Journal of Shandong University(Engineering Science). 2018, 48(6):  116-121.  doi:10.6040/j.issn.1672-3961.0.2018.142
    Abstract ( 488 )   HTML ( 4 )   PDF (4426KB) ( 143 )   Save
    Figures and Tables | References | Related Articles | Metrics

    To study the stiffener design discipline of large scale circular exhaust duct, finite element method was used to analyze the impacts on the strength, flexibility and buckling security from internal pressure load under different ash thickness, different number and specification stiffeners. The results indicated that the main stress on the circular duct lied on the contact zone of duct and support, the self-weight was the main factor of stress. For the large diameter duct, the stress under negative press was much higher than the condition of zero pressure and positive pressure. And for the large diameter duct, the stress may rapidly increase while the negative pressure rising, and the large diameter should not be used on negative duct. The ash thickness had smaller impact on the stress of duct, the buckling may less happen while the duct was thicker. The buckling factor of non-dusty condition should be checked on the design of considering dusty thickness.

    Motion control system design of multi-joint snake-like manipulator for nuclear environment
    Qiang ZHANG
    Journal of Shandong University(Engineering Science). 2018, 48(6):  122-131.  doi:10.6040/j.issn.1672-3961.0.2018.237
    Abstract ( 939 )   HTML ( 10 )   PDF (9206KB) ( 299 )   Save
    Figures and Tables | References | Related Articles | Metrics

    Based on the detection and maintenance work in the nuclear fusion reactor vessel as an example, a snake-like remote control manipulator with multi-joint series structure was presented, which adopted a composite control plan based on track push pull and adjustment with suspending arms. Aiming at the demand of the manipulator to carry out the whole vessel operation, the trajectory of the manipulator was simulated and analyzed. A motion control algorithm including path planning and trajectory control was designed, and a multi-axis coordination control system was established. The motion of the manipulator was tested inside the geometric simulation vessel environment, and the control accuracy of the joint rotation angle was evaluated. The gravity compensation module of the manipulator system was constructed, and the terminal positioning accuracy of the manipulator system was evaluated by the simulation of the flexible model. The test results verified the effectiveness of the motion control system.

    Polyaniline modified graphene layers/graphite plate electrode for supercapacitor
    Xiaodan WANG,Mingming GAO
    Journal of Shandong University(Engineering Science). 2018, 48(6):  132-136.  doi:10.6040/j.issn.1672-3961.0.2017.056
    Abstract ( 628 )   HTML ( 13 )   PDF (4672KB) ( 258 )   Save
    Figures and Tables | References | Related Articles | Metrics

    Polyaniline modified graphene layers/graphite plate (GL/GP/PANI) electrodes were prepared by electrochemical polymerization of aniline on graphene layers/graphite electrodes, which were obtained via in-situ formation of graphene layers on graphite plate by electrochemical exfoliation. The effect of polymerization cycle to specific capacitance of GL/GP/PANI was investigated. The surface features of the electrodes were characterized by SEM. The electrochemical performances of the electrodes were measured in 0.5 M H2SO4 electrolyte, including cyclic voltammetry, galvanostatic charge-discharge and electrochemical stability test. The results indicated that the PANI on the GL/GP electrodes had rodlike structure. At current density of 0.085 mA/cm2, the specific capacitance GL/GP/PANI reached 1 042.8 F/g. This study provided a new way to fabricate substrate electrode material for supercapacitor.