• Journal of Information Processing Systems
• /
• 제20권4호
• /
• pp.458-476
• /
• 2024

To address the issue of class imbalance in network traffic data, which affects the network intrusion detection performance, a combined framework using transformers is proposed. First, Tomek Links, SMOTE, and WGAN are used to preprocess the data to solve the class-imbalance problem. Second, the transformer is used to encode traffic data to extract the correlation between network traffic. Finally, a hybrid deep learning network model combining a bidirectional gated current unit and deep neural network is proposed, which is used to extract long-dependence features. A DNN is used to extract deep level features, and softmax is used to complete classification. Experiments were conducted on the NSLKDD, UNSWNB15, and CICIDS2017 datasets, and the detection accuracy rates of the proposed model were 99.72%, 84.86%, and 99.89% on three datasets, respectively. Compared with other relatively new deep-learning network models, it effectively improved the intrusion detection performance, thereby improving the communication security of network data.

• Korean Management Science Review
• /
• 제11권1호
• /
• pp.187-196
• /
• 1994

Time-Space-Time(T-S-T) switching network is modeled as a graceful degrading system. Call blocking probability is defined as a measure of performance. Several performance related measures are suggested under the presence of failure. An optimization model is proposed, which determines optimal values of system parameters of the switching network.

• Journal of KIISE:Information Networking
• /
• 제34권3호
• /
• pp.193-202
• /
• 2007

In this paper, a performance evaluation model of the Multistage Interconnection Network(MIN) with the multiple-buffered crossbar switches under Hot-spot environment is proposed and examined. Buffered switch technique is well known to solve the data collision problem of the MIN. The proposed evaluation model is developed by investigating the transfer patterns of data packets in a switch. The performance of the multiple-buffered $a{\times}a$ crossbar switch is analyzed. Steady state probability concept is used to simplify the analyzing processes. Two important parameters of the network performance, throughput and delay, are then evaluated. To validate the proposed analysis model, the simulation is carried out on a Baseline network that uses the multiple buffered crossbar switches. Less than 2% differences between analysis and simulation results are observed. It is also shown that the network performance is significantly improved when the small number of buffer spaces is given. However, the throughput elevation is getting reduced and network delay becomes increasing as more buffer spaces are added in a switch.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• 제12권11호
• /
• pp.4801-4811
• /
• 2011

The purpose of this study is to examine the factors influencing performance of internal, external environment and entrepreneurship in the social enterprise with network activity. This model tests various theoretical research hypotheses relating to performance of social enterprise, internal, external environment, entrepreneurship and network activity. The proposed model is analyzed to target social entrepreneurs. The results of hypothesis testing are as follows. First, organizational culture positively influence entrepreneurship. Second, environmental uncertainty positively influence entrepreneurship. Third, entrepreneurship positively influence managerial and social performance. Finally, organizational culture to entrepreneurship for high group of network activity is significantly larger than those for low group of network activity. In addition, environmental uncertainty to entrepreneurship for low group of network activity is significantly larger than those for high group of network activity.

• Phonetics and Speech Sciences
• /
• 제10권1호
• /
• pp.33-38
• /
• 2018

Although deep neural network-based acoustic models have greatly improved the performance of automatic speech recognition (ASR), reverberation still degrades the performance of distant speech recognition in indoor environments. In this paper, we adopt the DenseNet, which has shown great performance results in image classification tasks, to improve the performance of reverberant speech recognition. The DenseNet enables the deep convolutional neural network (CNN) to be effectively trained by concatenating feature maps in each convolutional layer. In addition, we extend the concept of multi-resolution CNN to multi-resolution DenseNet for robust speech recognition in reverberant environments. We evaluate the performance of reverberant speech recognition on the single-channel ASR task in reverberant voice enhancement and recognition benchmark (REVERB) challenge 2014. According to the experimental results, the DenseNet-based acoustic models show better performance than do the conventional CNN-based ones, and the multi-resolution DenseNet provides additional performance improvement.

• Journal of Digital Convergence
• /
• 제10권6호
• /
• pp.33-39
• /
• 2012

The purpose of this study is to investigate the effect of SNS(Social Network Service) on individual learning performance. To do this, we distribute and collect data by using a survey method. Research results suggest that online social networking engagement and acculturation have an effect on interaction quality with professors. Interaction quality with professors influences individual learning performance as well as collaborative learning. The conclusion and implications are discussed.

• Journal of information and communication convergence engineering
• /
• 제9권5호
• /
• pp.610-614
• /
• 2011

Nowadays Mobile Ad Hoc Networks (MANETs) are a very popular and emerging technology in the world. MANETs helps mobile nodes to communicate with each other anywhere without using infrastructure. For this purpose we need good routing protocols to establish the network between nodes because mobile nodes can change their topology very fast. Mobile node movements are very important features of the routing protocol. They can have a direct effect on the network performance. In this paper, we are going to discuss random walk and random waypoint mobility models and their effects on routing parameters. Previously, mobility models were used to evaluate network performance under the different routing protocols. Therefore, the network performance will be strongly modeled by the nature of the mobility pattern. The routing protocols must rearrange the changes of accurate routes within the order. Thus, the overheads of traffic routing updates are significantly high. For specific network protocols or applications, these mobility patterns have different impacts.

• The Journal of Information Systems
• /
• 제15권2호
• /
• pp.173-193
• /
• 2006

This study first conceptualizes and investigates five dimensions of SCM activities, such as organizational capacity, revitalization support collaboration, appraisal and responsibility, and information system activities. Second, this study develops three dimensions of a supply chain structure, and proposes a balanced scorecard (BSC) model in order to measure business performance. And then, this study establishes too research hypotheses as follows: H1. The SCM activities varies with the supply chain structure. H2. The management performance varies with the supply chain structure. In the questionnaire survey for empirical analysis, this study carefully selected 809 of companies in Korea. We conducted a survey by mail and collected 127 data. Out of 127 data we actually used 103 responses for statistical analysis. After conducting statistical analysis, we could find the results as followed: 1) The supply chain structure was classified into three networks through a clustering procedure, such as supply network, conversion network, and distribution network, and these networks were used to testify hypotheses. As a result the effect of SCM activities varies according to three networks and especially, the companies in distribution network were more active than those in supply and conversion networks doing SCM activities. 2) We may conclude that business performance varies with three networks, and distribution network achieves better performance than supply and conversion networks do.

• Journal of Satellite, Information and Communications
• /
• 제11권3호
• /
• pp.123-126
• /
• 2016

Energy efficiency in heterogeneous network is considered as one of the main issues when deploying the wireless sensor network. In heterogeneous network, the random distribution of initial energy at each node could lead to an instability of the network. Therefore, a resonable policy must be established in order to maintain the fairness in energy consumption and extend the working time of each node in the network. In this paper, we evaluate the performance of the distributed clustering protocol (DCP) in heterogeneous network on different scenarios. Simulation results are compared with the results of a LEACH protocol in a heterogeneous network. In addition, the performance of system in heterogeneous network are also compared with the homogeneous network to illustrate the effect of imbalance in the initial energy on the life time of each node in the system. The result illustrates that the DCP protocol demonstrates better performance than LEACH protocol in both the heterogeneous and the homogeneous networks.

• Journal of KIISE
• /
• 제44권7호
• /
• pp.668-673
• /
• 2017

Deep neural networks have shown remarkable performance in various fields of pattern recognition such as voice recognition, image recognition and object detection. However, underlying mechanisms of the network have not been fully revealed. In this paper, we focused on empirical analysis of the network parameters. The Faster R-CNN(region-based convolutional neural network) was used as a baseline network of our work and three important parameters were analyzed: the dropout ratio which prevents the overfitting of the neural network, the size of the anchor boxes and the activation function. We also compared the performance of dropout and batch normalization. The network performed favorably when the dropout ratio was 0.3 and the size of the anchor box had not shown notable relation to the performance of the network. The result showed that batch normalization can't entirely substitute the dropout method. The used leaky ReLU(rectified linear unit) with a negative domain slope of 0.02 showed comparably good performance.