Journals
  Publication Years
  Keywords
Search within results Open Search
Please wait a minute...
For Selected: Toggle Thumbnails
Extreme learning machine optimization based on hidden layer output matrix
SUN Haoyi, WANG Chuanmei, DING Yiming
Journal of Computer Applications    2021, 41 (9): 2481-2488.   DOI: 10.11772/j.issn.1001-9081.2020111791
Abstract346)      PDF (1706KB)(387)       Save
Aiming at the problem of the error existed from the hidden layer to the output layer of Extreme Learning Machine(ELM), it was found the analysis revealed that the error came from the process of solving the Moore-Penrose generalized inverse matrix H of the hidden layer output matrix H,that revaled the matrix H H was deviated from the identity matrix. The appropriate output matrix H was able to be selected according to the degree of deviation to obtain a smaller training error. According to the definitions of the generalized inverse matrix and auxiliary matrix,the target matrix H H and the error index L21-norm were firstly determined. Then,the experimental analysis showed that the L21-norm of H H was linearly related to the ELM error. Finally,Gaussian filtering was introduced to reduce the noise of the target matrix,which effectively reduced the L21-norm of the target matrix and the ELM error,achieving the purpose of optimizing the ELM algorithm.
Reference | Related Articles | Metrics
Prediction method of capacity data in telecom industry based on recurrent neural network
DING Yin, SANG Nan, LI Xiaoyu, WU Feizhou
Journal of Computer Applications    2021, 41 (8): 2373-2378.   DOI: 10.11772/j.issn.1001-9081.2020101677
Abstract510)      PDF (1094KB)(381)       Save
In the capacity prediction process of telecom operation and maintenance, there are problems of too many capacity indicators and deployed business classes. Most of the existing researches do not consider the difference of indicator data types, and use the same prediction method for all types of data, which results in both good and bad prediction effects. In order to improve the efficiency of indicator prediction, a classification method of data type was proposed, and the data types were divided into trend type, periodic type and irregular type. Aiming at the prediction of periodical data, a periodic capacity indicator prediction model based on Bi-directional Recurrent Neural Network (BiRNN), called BiRNN-BiLSTM-BI, was proposed. Firstly, In order to analyze the periodic characteristics of capacity data, a busy and idle distribution analysis algorithm was proposed. Secondly, a Recurrent Neural Network (RNN) model was built, which included a layer of BiRNN and a layer of Bi-directional Long Short-Term Memory network (BiLSTM). Finally, the output of BiRNN was optimized by the system's busy and idle distribution information. Experimental results compared with the best one among Holt-Winters, AutoRregressive Integrated Moving Average (ARIMA) model and Back Propagation (BP) neural network model show that, the proposed BiRNN-BiLSTM-BI model has the Mean Square Error (MSE) reduced by 15.16% and 45.67% on the unified log dataset and the distributed cache service dataset respectively, showing that the prediction accuracy is greatly improved.
Reference | Related Articles | Metrics
Multi-modal brain tumor segmentation method under same feature space
CHEN Hao, QIN Zhiguang, DING Yi
Journal of Computer Applications    2020, 40 (7): 2104-2109.   DOI: 10.11772/j.issn.1001-9081.2019122233
Abstract420)      PDF (874KB)(497)       Save
Glioma segmentation depends on multi-modal Magnetic Resonance Imaging (MRI) images. Convolutional Neural Network (CNN)-based segmentation algorithms are often trained and tested on fixed multi-modal images, which ignores the problem of missing or increasing of modal images. To solve this problem, a method mapping images of different modalities to the same feature space by CNN and using the features in the same feature space to segment tumors was proposed. Firstly, the features of different modalities were extracted through the same deep CNN. Then, the features of different modal images were concatenated, and passed through the fully connected layer to realize the feature fusion. Finally, the fused features were used to segment the brain tumor. The proposed model was trained and tested on the BRATS2015 dataset, and verified with the Dice coefficient. The experimental results show that, the proposed model can effectively alleviate the problem of data missing. At the same time, compared with multi-modal joint method, this model is more flexible, and can deal with the problem of modal data increasing.
Reference | Related Articles | Metrics
Application of improved GoogLeNet based on weak supervision in DR detection
DING Yingzi, DING Xiangqian, GUO Baoqi
Journal of Computer Applications    2019, 39 (8): 2484-2488.   DOI: 10.11772/j.issn.1001-9081.2019010225
Abstract482)      PDF (750KB)(266)       Save
To handle the issues of small sample size and multi-target detection in the hierarchical detection of diabetic retinopathy, a weakly supervised target detection network based on improved GoogLeNet was proposed. Firstly, the GoogLeNet network was improved, the last fully-connected layer of the network was removed and the position information of the detection target was retained. A global max pooling layer was added, and the sigmoid cross entropy was used as the objective function of training to obtain the feature map with multiple feature position information. Secondly, based on the weak supervision method, only the category label was used to train the network. Thirdly, a connected region algorithm was designed to calculate the boundary coordinate set of feature connected regions. Finally, the boundary box was used to locate the lesion in the image to be tested. Experimental results show that under the small sample condition, the accuracy of the improved model reaches 94%, which is improved by 10% compared with SSD (Single Shot mltibox Detector) algorithm. The improved model realizes end-to-end lesion recognition under small sample condition, and the high accuracy of the model ensures its application value in fundus screening.
Reference | Related Articles | Metrics
Dynamic allocation of virtual supplier resources based on cloud procurement platform
HUANG Li DING Yi YAO Jinyuan LIN Guolong
Journal of Computer Applications    2014, 34 (2): 377-381.  
Abstract590)      PDF (681KB)(435)       Save
This paper focused on the application of cloud computing technology to purchase link to form a cloud procurement platform, and to explore how to allocate the virtual machine with the virtual suppliers resources, so as to improve the satisfaction of customers. Firstly, this paper proposed the concept of cloud purchase platform, assuming that the virtual machine containing the suppliers of resources; secondly proposed the allocation processes of virtual machines which contained the virtual suppliers resources and modeling; then the Best Fit Decreasing (BFD) and Finder-tracker multi-swarm Particle Swarm Optimization (FTMPSO) were adopted to get the solution; finally the results of computing were analyzed. In the BFD algorithm, the priority of each of the three attributes met different preferences of the customer's requirements. Using FTMPSO algorithm to allocate virtual suppliers resources got higher satisfaction of customer than using BFD to allocated virtual suppliers resources.
Related Articles | Metrics
Empirical analysis of symmetry degree for micro-blog social network
KANG Zedong YU Jinghu DING Yiming
Journal of Computer Applications    2014, 34 (12): 3405-3408.  
Abstract215)      PDF (811KB)(617)       Save

While Twitter and Sina micro-blogs abundant registered users formed a social network of focusing relationship, by using the degree of symmetry its change regulation with the scale of the social circle was studied. Firstly, based on the collection of 1000000 focusing relationships among the Sina micro-blog users and 236 Twitter users as well as their focusing relationships, the initial social network was established. Here focus lied on the connected sub-networks which had obvious symmetrical connects, then the elimination method was applied to obtain these conclusions: The major factors that affect the symmetry of the maximum connected sub-networks are those who are called big V users and negligible users. After that, comparative analysis method was used to find out that the sub-network consisted of the big V users in Twitter has a stronger symmetry. Finally, the difference between these two kinds of micro-blogs was figured out in terms of functional localization. Through the researches on the symmetry of all connected sub-networks within the initial network, the result shows that when the scale of a public social circle decreases, the corresponding symmetry becomes stronger.

Reference | Related Articles | Metrics
Multiple time series autoregressive method based on support vector regression
ZHANG Wei LIU Xian-hui DING Yi SHI De-ming
Journal of Computer Applications    2012, 32 (09): 2508-2511.   DOI: 10.3724/SP.J.1087.2012.02508
Abstract1050)      PDF (722KB)(607)       Save
Energy consumption time series involves a variety of energy and the relationship between different energy is complicated. Most existing consumption methods make prediction through multiple independent single time series respectively, which ignores dependencies between multiple time series. In order to take full advantage of the association between multiple time series and improve prediction accuracy, the vector-valued autoregressive method and multi-task autoregressive method based on Support Vector Regression (SVR) machines were proposed for multiple time series forecast according to vector-valued function learning and multi-task learning theory. The experimental results with energy consumption of coking process verify that multiple time series autoregressive models based on the proposed methods show better prediction performance.
Reference | Related Articles | Metrics