• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.22 no.10
• /
• pp.1300-1306
• /
• 2018

The purpose of this study is to see possibility of Char2Vec as alternative of Word2Vec that most famous word embedding model in Sentence Similarity Measure Problem by Deep-Learning. In experiment, we used the Siamese Ma-LSTM recurrent neural network architecture for measure similarity two random sentences. Siamese Ma-LSTM model was implemented with tensorflow. We train each model with 200 epoch on gpu environment and it took about 20 hours. Then we compared Word2Vec based model training result with Char2Vec based model training result. as a result, model of based with Char2Vec that initialized random weight record 75.1% validation dataset accuracy and model of based with Word2Vec that pretrained with 3 million words and phrase record 71.6% validation dataset accuracy. so Char2Vec is suitable alternate of Word2Vec to optimize high system memory requirements problem.

• Communications for Statistical Applications and Methods
• /
• v.15 no.1
• /
• pp.105-115
• /
• 2008

Judgement post-stratification is a new sampling method developed by MacEachern et al. (2004). This article suggests that the judgement post-stratification method can be a good alternative for the simple random sampling when analyzing real-world environmental data. It becomes an important task to accurately measure the output of a recycling facility since the EPR (Extended Producer Responsibility) system takes effect on 2003. However, the total weight of materials processed in the recycling facility may not be a proper measure because the materials are frequently mingled with other non-recycling materials. Therefore, it is necessary to estimate the mixture ratio of non-recycling materials among the total materials admitted in the facility. Unfortunately, the size of sample in a recycling facility is restricted due to the inconvenience of sampling procedure such as safety, odor, time and classification of non-recycling materials. In this article, we showed the relative efficiency of the judgement post-stratification method over the simple random sampling method for equal sample sizes using Monte Carlo simulation. Furthermore, we applied the judgement post-stratification method on the 2004 recycling data and showed that it can replace the simple random sampling even with smaller observations.

• Journal of Korea Multimedia Society
• /
• v.13 no.8
• /
• pp.1128-1141
• /
• 2010

This paper proposes a method of moving object detection in crowded scene using clustered trajectory. Unlike previous appearance based approaches, the proposed method employes motion information only to isolate moving objects. In the proposed method, feature points are extracted from input frames first and then feature tracking is followed to create feature trajectories. Based on an assumption that feature points originated from the same objects shows similar motion as the object moves, the proposed method detects moving objects by clustering trajectories of similar motions. For this purpose an energy function based on spatial proximity, motion coherence, and temporal continuity is defined to measure the similarity between two trajectories and the clustering is achieved by minimizing the energy function in CRFs (conditional random fields). Compared to previous methods, which are unable to separate falsely merged trajectories during the clustering process, the proposed method is able to rearrange the falsely merged trajectories during iteration because the clustering is solved my energy minimization in CRFs. Experiment results with three different crowded scenes show about 94% detection rate with 7% false alarm rate.

• Journal of KIISE:Information Networking
• /
• v.28 no.4
• /
• pp.517-526
• /
• 2001

In order to reduce the increasing packet loss rates caused by an exponential increase in network traffic, the IETF(Internet Engineering Task Force) is considering the deployment of active queue management techniques such as RED(Random Early Detection). While active queue management in routers and gateways can potentially reduce total packet loss rates in the Internet, this paper has demonstrated the inherent weakness of current techniques and shows that they are ineffective in preventing high loss rates. The inherent problem with these queue management algorithms is that they all use queue lengths as the indicator of the severity of congestion. In this paper, in order to solve this problem, a new active queue management algorithm called MRED(Modified Random Early Detection) is proposed. MRED computes the packet drop probability based on our heuristic method rather than the simple method used in RED. Using simulation, MRED is shown to perform better than existing queue management schemes. To analyze the performance, we also measure throughput of traffics under the FIFO control, and compared the performance with that of this MRED system.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.26 no.11
• /
• pp.999-1011
• /
• 2015

In modern electronic warfare, RADAR is under constant threat of ECM(Electronic Counter Measures) signals from nearby jammers. The conventional linear frequency modulated(Linear-FM) waveform is easy to be intercepted to estimate its signal parameters due to its periodical phase transition. Recently, APCN(Advanced Pulse Compression Noise) waveform using random amplitude and phase transition was proposed for LPI(Low probability of Intercept). But random phase code signals such as APCN waveform tend to be sensitive to Doppler frequency shift and result in performance degradation during moving target detection. In this paper, random phase and code rate transition based radar waveform(RPCR) is proposed for Doppler tolerance improvement. Time frequency analysis is carried out through ambiguity analysis to validate the improved Doppler tolerance of RPCR waveform. As a means to measure the vulnerability of the proposed RPCR waveform against LPI, WHT(Wigner-Hough Transform) is adopted to analyze and estimate signal parameters for ECCM(Electronic Counter Counter Measures) application.

• The Journal of Korean Physical Therapy
• /
• v.27 no.4
• /
• pp.240-245
• /
• 2015

Purpose: The purpose of this study was to compare the blocked practice and random practice of task-oriented training in patients with chronic stroke to determine the effect of lower extremity muscle activity and balance ability. Methods: The thirty participants were randomly assigned to either the block practice group (BP) group or the random practice group (RP) and received the training three times per week, 30 minutes per day, for six weeks. Surface electromyography was used for measurement of lower extremity muscle activity. Static balance was to measured the stability index (SI) and weight distribution index (WDI) using the Tetrax. The four square step test (FSST) was used to measure dynamic balance. The paired t-test was used for determination of differences before and after intervention, and the independent t-test was used for determination of differences between groups. Results: Lower extremity muscle activity, RA and GCM was improved in the RP group after intervention and between groups. TA was significantly improved in the RP group compared with the BP group. In comparison of before and after interventions, SI was reduced in BP and RP. WDI in OS was reduced in comparison of BP and RP before and after intervention. CS was reduced in BP and RP. The OS and CS was improved in RP compared with BP. In comparison of before and after intervention, FSST was improved in BP and RP. Conclusion: Task-oriented training methods using random practice was found to be effective in promoting lower extremity muscle activity and balance ability in chronic stroke patients.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.12 no.6
• /
• pp.543-548
• /
• 2002

It is well known that FCM is an indispensible tool for fuzzy clustering. The problems of using FCM are 1) it is sensitive to the initial random membership functions and 2) FCM inherently requires the number of clusters. Hence we need to run FCM algorithms with an appropriate validity measure until we find a suitable number of clusters. In this paper, we suggest the Majority-Voting FCM with implied validity measure. With this algorithm, we can solve the aforementioned problems. The working simulation results are provided. The contributions are 1) MV-FCM algorithm and 2) its definitive capability of being an excellent validity measure.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.19 no.11
• /
• pp.2160-2169
• /
• 1994

Main components of phase time error of synchronous network are flicker noise and random walk noise. This paper describe computer simulation of clock error characterized by a statistical model recommended as a standard measure. Flicker noise sequences are generated from white noise sequences by means of a algorithm developed by Barnes. Random-walk noise sequence are obtained by integration of a white noise sequence. Especially for flicker noise, relation between stage number N, time constant ratio K and bandwidth of flicker noise generated was defined by using some examples.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.3
• /
• pp.27-32
• /
• 2020

In this paper, we propose an algorithm to measure the accumulated status of scrap boxes where metal scraps are accumulated. The accumulated status measuring is defined as a multi-class classification problem, and the method with deep learning classify the accumulated status using only the scrap box image. The learning was conducted by the Transfer Learning method, and the deep learning model was NASNet-A. In order to improve the accuracy of the model, we combined the Random Forest classifier with the trained NASNet-A and improved the model through post-processing. Testing with 4,195 data collected in the field showed 55% accuracy when only NASNet-A was applied, and the proposed method, NASNet with Random Forest, improved the accuracy by 88%.

• Journal of the Korean Data and Information Science Society
• /
• v.11 no.2
• /
• pp.189-199
• /
• 2000

A major deficiency of laboratory testing of rock structure is that the structures are limited in size and therefore present a very small and highly selective sample of the rock mass from which were removed. In a typical engineering project, the samples tested in the laboratory represent only a very small fraction of one percent of the volume of the rock mass. In this paper, we calculate the representative orientation of the resultant vector, the measure of the degree of clustering, the volume of rock mass, the trace length of discontinuity spacing under underlying distributions. And we generate the random fracture networks using real data. We propose the calculating the trace length.