Top Read

Published in last 1 year |

In last 2 years |

In last 3 years |

All

Please wait a minute...


For Selected: Download Citations EndNote Ris BibTeX Toggle Thumbnails

Select

Product Competitiveness Analysis Method Based on Online Reviews TIAN Jin-ni， YOU Tian-hui， YUAN Yuan Journal of Northeastern University(Natural Science)    2021, 42 (10): 1498-1505.   DOI: 10.12068/j.issn.1005-3026.2021.10.018 Abstract （3988）   HTML （13）    PDF（pc） （590KB）（684）       Knowledge map    Save With the development of e-commerce， online reviews have been regarded as important data resources for enterprises to analyze the competitiveness of their products. Product feature dimensions that consumers pay most attention to were extracted from review texts. The sentiment analysis of review texts was performed based on the sentiment lexicon method to obtain the feature-sentiment score pairs. The opportunity scores of feature dimensions were calculated by the importance degree and satisfaction degree， and an opportunity landscape diagram was drawn to analyze the competitiveness of the target product itself. Feature dimension satisfaction degree was calculated to compare the competitive advantages and disadvantages between the target product and its competitive products. Competitiveness improvement strategies of the target product were given by considering the competitiveness analysis results from two perspectives comprehensively. The online reviews provided by Zol.com.cn were used to conduct experiments to illustrate the feasibility and effectiveness of the proposed method. Reference | Related Articles | Metrics | Comments（0）

Select

Feature Selection Algorithm Based on LightGBM LI Zhan-shan， YAO Xin， LIU Zhao-geng， ZHANG Jia-chen Journal of Northeastern University(Natural Science)    2021, 42 (12): 1688-1695.   DOI: 10.12068/j.issn.1005-3026.2021.12.003 Abstract （3913）   HTML （98）    PDF（pc） （419KB）（8073）       Knowledge map    Save In order to solve the shortcomings of the following two types of feature selection algorithms， filtering and wrapping based on evolutionary learning， a new wrapping feature selection algorithm LGBFS(LightGBM feature selection) was proposed. First， LightGBM was introduced to construct an gradient boosting tree model for the original features and measure the importance of features; then the proposed LR sequential forward search strategy LRSFFS was combined to select features; finally， the proposed algorithm was compared with nine algorithms in 21 standard datasets. The results show that 16 of the 21 standard data sets of LGBFS have achieved the best classification accuracy， and 18 standard data sets have achieved the best dimensionality reduction rate and the best CPU running time. In addition， time complexity analysis and significance test were carried out. The test shows that LGBFS is significantly different from the six comparison algorithms， and it also shows that LGBFS can balance the calculation efficiency and classification accuracy of feature subsets. Reference | Related Articles | Metrics | Comments（0）

Select

Entity Alignment Algorithm for Knowledge Graph of Representation Learning ZHU Ji-zhao， QIAO Jian-zhong， LIN Shu-kuan Journal of Northeastern University Natural Science    2018, 39 (11): 1535-1539.   DOI: 10.12068/j.issn.1005-3026.2018.11.004 Abstract （3349）   HTML    PDF（pc） （394KB）（2617）       Knowledge map    Save A novel supervised method for knowledge graph entity alignment based on representation learning was proposed， which is different from the existing methods due to the similarity of structural information or attributive characters. First， the method automatically learns the semantic representations for the entities and relations of a knowledge graph in a low-dimensional vector space was proposed， and these embeddings contain the intrinsically structural information of a knowledge graph and the attributive features of entities. Afterwards， taking the manually aligned entity pairs as prior knowledge， the cross-KG mapping relationship between entities could be learned， which will be used for predicting entity alignment. Experiments conducted on real datasets demonstrated that our method can effectively improve the precision of knowledge graph entity alignment while keeping a high F1 score， when compared with the feature matching based method SiGMa. Reference | Related Articles | Metrics | Comments（0）

Select

Effect of ZrO2 content on the density of nNaF·AlF3-Al2O3- CaF2 molten salt system Bao, Morigen-Gaowa (1); Wang, Zhao-Wen (1); Gao, Bing-Liang (1); Shi, Zhong-Ning (1) Journal of Northeastern University    2012, 33 (6): 866-869.   DOI: - Abstract （3265）   HTML    PDF（pc） （439KB）（971）       Knowledge map    Save The density of nNaF·AlF3-Al2O3- CaF2-ZrO2 molten salt system with different molar ratio of NaF to AlF3 and ZrO2 contents was determined based on Archimedes law, and the effects of temperature and ZrO2 content were investigated. The results show that the density decreases with the increasing of temperature and the decreasing of molar ratio of NaF to AlF3. At 980°C, there is a good linear relationship between the density of 2.4NaF·AlF3-3wt%Al2O3- 3wt%CaF2 molten salt system and the ZrO2 content of 0~5wt%, which can be expressed as ρt(g·cm-3) = 0.02844·w(ZrO2)+2.045 (r=0.9992). Within the present study range, the electrolyte density increases approximately by 0.02 g/cm3 with 1wt% ZrO2 addition. Reference | Related Articles | Metrics | Comments（0）

Select

Traffic Target Detection in Complex Scenes Based on Improved YOLOv5 Algorithm GU De-ying， LUO Yu-lun， LI Wen-chao Journal of Northeastern University(Natural Science)    2022, 43 (8): 1073-1079.   DOI: 10.12068/j.issn.1005-3026.2022.08.002 Abstract （3003）   HTML （92）    PDF（pc） （2504KB）（1222）       Knowledge map    Save Real-time target detection in traffic scenarios is the prerequisite of electronic monitoring， automatic driving， and other functions. In view of the low detection efficiency of existing target detection algorithms and the low accuracy of most light target detection algorithms， which are easy to obtain wrong or insufficient target detection， this paper adopts the improved YOLOv5 target detection algorithm for model training， and the pseudo-label strategy for training process optimization. Then， the KITTI traffic target dataset tags are merged into three categories for model training and testing. Through the experimental comparison， the improved YOLOv5 model in this paper achieves 92.5% mAP in all categories， which is 3% higher than the original YOLOv5 training model. Finally， the three categories of the trained models are deployed on the Jetson Nano embedded platform for inference testing， and TensorRT is used to accelerate inference. The average inference time per frame of image is 77ms， which meets the goal of real-time detection. Reference | Related Articles | Metrics | Comments（0）

Select

Trajectory Tracking Control of Unmanned Vehicles Based on Adaptive MPC LIANG Zhong-chao， ZHANG Huan， ZHAO Jing， WANG Yong-fu Journal of Northeastern University Natural Science    2020, 41 (6): 835-840.   DOI: 10.12068/j.issn.1005-3026.2020.06.013 Abstract （2845）   HTML    PDF（pc） （562KB）（3279）       Knowledge map    Save According to the principle of adaptive MPC(model predictive control)， a trajectory tracking control strategy for unmanned vehicles is designed. Based on the vehicle dynamics model， the trajectory tracking controller is established， and the objective function and related constraints are designed. The adaptive MPC control algorithm is used for solution. At each control point of operation， the Kalman state estimator correlation gain coefficient matrix and the state of the controller are continuously updated to adapt to the current working environment of the unmanned vehicle， thereby compensating for the nonlinearity of the vehicle and the effect of state measurement noise. The simulation model is built in MATLAB and verified by simulation. It is concluded that the adaptive MPC has better control precision and robustness for the trajectory tracking of unmanned vehicles. The effectiveness of the algorithm in the trajectory tracking control layer is verified. In addition， it provides a reference for the research of future trajectory tracking control. Reference | Related Articles | Metrics | Comments（0）

Select

Extraction and Adsorption Properties of Extracellular Polymeric Substance XU Xiao-hui， WEI De-zhou， ZHANG Lan-he Journal of Northeastern University(Natural Science)    2021, 42 (10): 1467-1474.   DOI: 10.12068/j.issn.1005-3026.2021.10.014 Abstract （2784）   HTML （61）    PDF（pc） （1430KB）（1146）       Knowledge map    Save In order to quickly and efficiently extract extracellular polymeric substances(EPS)from activated sludge， four methods were used to extract tightly bound EPS(TB-EPS). The advantages and disadvantages of different extraction methods were compared， the effect of TB-EPS on the adsorption performance of Mn2+was investigated， and the adsorption mechanism of TB-EPS was studied. The results show that the heating method is simple and efficient with no damage to cell structure.When adsorbing heavy metal Mn2+， the maximum adsorption rate of TB-EPS extracted by heating method is 53.8%. The adsorption process conforms to the second-order reaction kinetic characteristics and the Langmuir adsorption isotherm equation， which is mainly the chemical adsorption. The main adsorption mechanism of TB-EPS is the interaction of hydrophobic tryptophan and tyrosin. Reference | Related Articles | Metrics | Comments（0）

Select

Object Detecting and Tracking Algorithm Based on Optical Flow XIAO Jun， ZHU Shi-peng， HUANG Hang， XIE Ya-nan Journal of Northeastern University Natural Science    2016, 37 (6): 770-774.   DOI: 10.12068/j.issn.1005-3026.2016.06.003 Abstract （2701）   HTML    PDF（pc） （318KB）（2792）       Knowledge map    Save Harris corner points were adopted as tracking objects， and scale space was introduced into corner point detection in order to extract Harris corner points in feature scale. Then curvature was computed to filter out false corners and enhance adaptability to scale change. Optical flow method was adopted for the tracking algorithm based on image pyramid， in which the optical flow iteratively was computed. And the tracking algorithm based on the optical flow error was proposed. That is， the trajectory error in the same frame with different time flow was used to evaluate the tracking situation. In this way， tracking failure was avoided when the tracking object is hidden， disappears or textural features change. Experimental results of different video sequences show that the proposed optical flow tracking algorithm based on improved corner extraction and image pyramid has better tracking performances. The feature points could be filtered effectively that lead to tracking failure with the introduction of optical flow error method， and the object positions are estimated accurately. Reference | Related Articles | Metrics | Comments（0）

Select

Heavy Metal Pollution and Potential Risk Assessment of Soil in Fushun Western Open-Pit Mining Area JIANG Bin-hui， ZHANG Bo， WANG Xue-feng， JIANG Qi Journal of Northeastern University Natural Science    2020, 41 (4): 568-574.   DOI: 10.12068/j.issn.1005-3026.2020.04.020 Abstract （2669）   HTML    PDF（pc） （558KB）（1528）       Knowledge map    Save By measuring the heavy metal content in the typical area of the mining area， it is found that Cd is the characteristic element of the heavy metal pollution in the mining area. The pollution degree and the potential ecological risk caused by the heavy mental pollution were evaluated by using the single factor pollution index method， the Nemero pollution index method and the potential hazard index method respectively. The results showed that the average values of single element pollution index and Nemero pollution index of Cd are 4.06 and 5.32， respectively， both indicate a heavy pollution. The Nemero pollution index of the three ecological restoration areas are 2.07~4.57， among which the south natural ecological restoration(SNER) area is heavy pollution and the other areas are moderate pollution. The potential ecological risk index of heavy metals in the mining area is 98.62~224.55， the SNER area reaches the medium hazard level， and the remaining areas are evaluated as mild ecological hazards. Reference | Related Articles | Metrics | Comments（0）

Select

Experimental Study on Laser Heat-Assisted Grinding Quartz Glass YU Tian-biao， LIU Chuan-meng， YU Kui-dong， WANG Wan-shan Journal of Northeastern University Natural Science    2019, 40 (10): 1467-1473.   DOI: 10.12068/j.issn.1005-3026.2019.10.018 Abstract （2595）   HTML    PDF（pc） （2190KB）（1240）       Knowledge map    Save The effects of process parameters on the surface roughness， surface morphology and wear of the grinding wheel were studied by the orthogonal test after quartz glass was treated by laser heat assisted grinding. The results showed that laser heat-assisted grinding can improve the critical grinding depth， the surface grinding quality and work efficiency of quartz glass. The laser power has the greatest influence on the roughness of laser-assisted grinding， but there is no linear relationship. The optimal laser power is 175W， and the corresponding roughness is 0.262. With the aid of laser， the brittleness of glass decreases and the plasticity increases during the experiment， which realizes the plastic domain grinding of quartz glass， reduces the wear of the grinding wheel and decreases the peeling pit of the grinding surface. Reference | Related Articles | Metrics | Comments（0）

Select

Process Mineralogical Characteristics and Ore Value of Typical Vanadium Titanium Magnetite in China YU Hong-dong， WANG Li-na， QU Jing-kui， QI Tao Journal of Northeastern University Natural Science    2020, 41 (2): 275-281.   DOI: 10.12068/j.issn.1005-3026.2020.02.022 Abstract （2514）   HTML    PDF（pc） （1843KB）（2318）       Knowledge map    Save The process mineralogy of vanadium-titanium magnetite from Panzhihua， Chengde and Chaoyang is studied， and the value of the ore is calculated according to the spot price of iron concentrate， titanium concentrate and the theoretical mineral processing index. The results show that the grade of magnetic iron is 13.26%， 12.84% and 1.15%， the grade of TiO2 is 10.98%， 4.53% and 2.02%， and the grade of V2O5 is 0.22%， 0.23% and 0.055%， respectively， in vanadium- titanomagnetite from Panzhihua， Chengde and Chaoyang. The mineral contents and mineral compositions of titanomagnetite and ilmenite in vanadium-titanomagnetite of different mineral areas are significantly different. The grade of iron is 59.21%， 66.46% and 55.74%， the grade of TiO2 is 12.35%， 4.87% and 16.42%， and the grade of V2O5 is 0.928%， 1.142% and 2.07%， respectively， in iron concentrates from Panzhihua， Chengde and Chaoyang. Panzhihua ore has the highest value， followed by Chengde， while Chaoyang ore is the lowest. Reference | Related Articles | Metrics | Comments（0）

Select

Pavement Crack Detection Algorithm Based on Subregion and Multiscale Analysis LU Ziwei， WU Chengdong， CHEN Dongyue， SHANG Shibo Journal of Northeastern University Natural Science    2014, 35 (5): 622-625.   DOI: 10.12068/j.issn.1005-3026.2014.05.004 Abstract （2491）   HTML    PDF（pc） （676KB）（1873）       Knowledge map    Save In order to improve the accuracy of highway pavement crack detection， a new type of surface defects detection algorithm was proposed based on subregion and multiscale analysis. Gray， entropy and texture features distribution information were extracted in cracks and the surrounding areas from different scales of images. Feature vectors of parameters containing the direction trend and bending degree were acquired， and the crack location was detected through learning the support vector machine (SVM) and judging the eigenvector. Experimental results demonstrated that the resistance to noise， versatility and detection accuracy were improved effectively by the proposed algorithm in comparison to the other pavement cracks detection algorithms. The ideal crack detection effect was achieved， and the requirements of highway quality inspection were met effectively. Reference | Related Articles | Metrics | Comments（0）

Select

Analysis on Mechanical Properties of the Steel Plate Shear Wall with Fully Bolted and Welded Connections MU Zai-gen， DA Hai， YANG Yu-qing， WANG Zhe Journal of Northeastern University(Natural Science)    2021, 42 (3): 428-435.   DOI: 10.12068/j.issn.1005-3026.2021.03.019 Abstract （2477）   HTML （12）    PDF（pc） （5074KB）（435）       Knowledge map    Save In order to avoid the residual stress caused by welding， the upper and lower steel plates are connected by the end-plate with bolts to form a fully bolted steel plate shear wall (SPSW) with better ductility and energy dissipation capacity. The models of fully bolted SPSW and welded SPSW were established and analyzed by the finite element method. The load-displacement curve， lateral bearing capacity， energy dissipation capacity， stiffness degradation， and bearing capacity degradation of the two kinds of steel plate shear walls under unidirectional push-over and reciprocating load were compared. The results show that， under the action of unidirectional push-over， the lateral bearing capacity of the SPSW with full bolted is close to that of the traditional form， but the initial stiffness is 11% lower than that of the traditional form. Under the reciprocating load， the SPSW with full bolted shows a better overall energy dissipation capacity， and the stiffness degrades slower. This form of SPSW is also convenient for production and installation， and worth to extensive promotion and application in the future engineering application. Reference | Related Articles | Metrics | Comments（0）

Select

Numerical Simulation of Dynamic Response of Surrounding Rock Anchored by NPR Bolt/Cable GUO Long-ji， HE Man-chao， QU Ding-jun， TAO Zhi-gang Journal of Northeastern University(Natural Science)    2022, 43 (8): 1159-1167.   DOI: 10.12068/j.issn.1005-3026.2022.08.013 Abstract （2420）   HTML （6）    PDF（pc） （3435KB）（479）       Knowledge map    Save In this paper， a coal mine is used as the research region to study the problems caused by deep mining， such as rock burst， large deformation and failure of conventional bolt/cable. Based on the large deformation failure mechanism of high in-situ stress soft rock， the FLAC3D-PFC3D coupling modeling method was used to carry out the dynamic numerical simulation study of NPR (negative Poisson’s ratio) bolt/cable for the first time. Numerical models of two support schemes of PR and NPR bolt/cable were established， and the stress and displacement of the surrounding rock of the chamber and the bolt/cable under dynamic loads were compared and analyzed. The results show that under the impact load of 25MPa， 21 of the 25 PR support bolts/cables failed， the surrounding rock was damaged in a large area， the coupling wall was severely deformed， and the chamber presented inner convex arc-shaped deformation and failure. Under the same conditions， none of the NPR bolts/cables failed， and the stability of the surrounding rock of the chamber remained good. Reference | Related Articles | Metrics | Comments（0）

Select

Characteristics of Phase Change Materials and Analysis of Thermal Properties WUSIMAN Kuerbanjiang， DAI Xiao-ye， SHI Lin Journal of Northeastern University(Natural Science)    2023, 44 (4): 510-516.   DOI: 10.12068/j.issn.1005-3026.2023.04.007 Abstract （2370）   HTML （28）    PDF（pc） （802KB）（1585）       Knowledge map    Save Phase change materials (PCMs) are the core media of energy storage for latent heat storage. Thus， screening of appropriate PCMs needs to take into consideration a variety of factors， whereas phase transition temperature， energy storage density and thermal conductivity are the most critical factors during assessment of feasibility for PCMs. As evident by summarizing and comparing the characteristics of different categories of typical solid-liquid PCMs， multi-high-performance compatible composite materials is the development direction for PCMs’ applications. Finally， based on the theoretical analysis， it concluded as that within the range of specified working temperature difference of 50℃， the heat storage density per unit mass of common PCMs is no longer correlated with it’s phase transition temperature when latent heat is ≥100kJ/kg， and the higher of phase transition temperature of PCMs is， the wider range of the effective output temperature can be obtained by the system， and more flexible output of system becomes. The larger the latent heat of PCMs is， the more stable the heat discharging efficiency of the system is while varying the effective temperature. Therefore， within the temperature difference of 50℃， the thermal energy storage system gains rather superior output performance with selection of PCMs under considering of it’s higher latent heat， competitive average solid-liquid specific heat and higher phase transition temperature. Reference | Related Articles | Metrics | Comments（0）

Select

Experimental Study on Ignition and Explosion of Carbon Material Dust LI Gang， LIU Zong-yang， CHANG Wei-da， ZHANG Xiao-yu Journal of Northeastern University(Natural Science)    2023, 44 (2): 298-304.   DOI: 10.12068/j.issn.1005-3026.2023.02.019 Abstract （2354）   HTML （34）    PDF（pc） （683KB）（1397）       Knowledge map    Save In order to clarify the dust explosion hazard of carbon materials， the explosiveness of 6 kinds of carbon dust materials was studied according to 3 standards. According to VDI 2263-1—1990， the low-carbon earthy graphite， the pigment carbon black and the conductive carbon black are explosive dust， low-carbon flake graphite， high-carbon flake graphite， and high-carbon earthy graphite are non-explosive dust. According to ASTM E 1226 and GB/T 16425—2018， all the 6 kinds of samples are explosive. The sensitivity characteristics of various samples were tested according to the existing test standards， the 6 types of powder dust clouds was not fired，but the dust layer of conducting charcoal black and pigment charcoal black fired at 450℃. By using the TG-DSC method ， it is found that high-carbon flake graphite， high-carbon earthy graphite and conductive carbon black experienced an oxidative exothermic reaction in the air atmosphere， with a starting temperature of 770， 680， 500℃ respectively. The explosiveness of graphite and carbon black depend on the ignition energy， and explosion will occur if it exceeds 10kJ. Although the fire sensitivity of graphite and carbon black dust is very low， there still exists fire risk. Reference | Related Articles | Metrics | Comments（0）

Select

Bearing Remaining Useful Life Prediction Method Based on Transfer Learning WANG Xin-gang， HAN Kai-zhong， WANG Chao， LI Lin Journal of Northeastern University(Natural Science)    2021, 42 (5): 665-672.   DOI: 10.12068/j.issn.1005-3026.2021.05.009 Abstract （2330）   HTML （75）    PDF（pc） （1780KB）（1990）       Knowledge map    Save To address the problem that most bearing remaining useful life (RUL) prediction methods based on artificial intelligence cannot well predict bearing RUL under different working conditions， a transfer learning method was proposed to predict bearing RUL under different working conditions. Fourier transform was applied to the raw vibration signals of the bearing to obtain the frequency-domain signals， and convolutional neural network (CNN) and long short-term memory network(LSTM) were used to extract the features between data of the bearing′s frequency-domain signals and mine temporal information. The method of combining global and local domain adaption was adopted to reduce the distribution differences of the bearing data under different working conditions. The effectiveness of the method was verified by the existing bearing data. Compared with the traditional deep learning models， the proposed method improves the accuracy of bearing RUL prediction under different working conditions. Reference | Related Articles | Metrics | Comments（0）

Select

Two-Stage U-Net Coronary Artery Segmentation Based on CTA Images WANG Lu， YANG Xiao-fan， WANG Qian-jin， XU Li-sheng Journal of Northeastern University(Natural Science)    2022, 43 (6): 792-800.   DOI: 10.12068/j.issn.1005-3026.2022.06.005 Abstract （2324）   HTML （63）    PDF（pc） （1747KB）（934）       Knowledge map    Save There are many challenges in coronary artery segmentation based on CT angiography(computed tomography angiography)images， such as the complex structure of the coronary artery， the severely unbalanced distribution of the coronary artery and the background， and the susceptibility to interference from the coronary veins and other tissues during segmentation. A two-stage coronary artery segmentation method was proposed. In the first stage， the 3D DRU-Net with dense feature extraction and residual feature correction capabilities was used to ensure the segmentation recall rate. In the second stage， a 2D dual encoding multi-feature fusion U-Net(2D DEMFU-Net)was developed to perform fine segmentation， in which features were extracted from the original image slices and the segmentation results produced by the first stage respectively， and the multi-level semantic features from these two branches were fused using dense skip connections， further improving the accuracy of segmentation. Experimental results show that the proposed two-stage segmentation algorithm outperforms 3D U-Net by 3.83%， 5.31% and 2.23% in terms of Dice similarity coefficient， recall rate and precision， respectively. Reference | Related Articles | Metrics | Comments（0）

Select

Heart rate extraction algorithm based on ballistocardiogram Wang, Chun-Wu (1); Wang, Xu (1); Long, Zhe (1); Zhang, Ke-Xin (1) Journal of Northeastern University    2012, 33 (8): 1103-1106.   DOI: - Abstract （2309）   HTML    PDF（pc） （1465KB）（2541）       Knowledge map    Save A novel algorithm was presented to extract heart rate from ballistocardiogram (BCG) signals more accurately. Original signals are pretreated by polynomial fitting and wavelet transform. Then priori heart rate is acquired through the spectrum analysis of unipolar signals and a local heartbeat cycle is divided into several subintervals based on the priori heart rate. Finally, the heart rate is obtained by threshold processing and local maximum value positioning. The BCG signals from 30 normal subjects were acquired, with their single-channel electrocardiogram (ECG) signals gathered synchronously. Compared the heart rate extracted from the BCG signals with that from ECG signals, the maximum relative error is only 1.78%, which verified the accuracy and feasibility of the proposed algorithm. Reference | Related Articles | Metrics | Comments（0）

Select

An Effective Proportional Fairness Algorithm LIU Fulai， QIU Huizhen， ZHOU Qingping， DU Ruiyan Journal of Northeastern University Natural Science    2014, 35 (11): 1535-1538.   DOI: 10.12068/j.issn.1005-3026.2014.11.004 Abstract （2275）   HTML    PDF（pc） （455KB）（1004）       Knowledge map    Save To solve the spectrum resource allocation problem of multiuser in cognitive radio system based on OFDM， a bit and power allocation algorithm was proposed. Firstly， the transmit power of secondary users was constrained to guarantee the communication quality of service (QoS) of secondary user. Secondly， the twostep method and the proportion gene were used to avoid the disequilibrium of spectrum resource allocation. Finally， the algorithm was simulated and the results of the allocation about secondary users showed that the presented method could not only guarantee each user’s QoS but also increase the channel capacity. Reference | Related Articles | Metrics | Comments（0）