• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권12호
• /
• pp.3889-3903
• /
• 2022

To address the problem of low detection accuracy due to training noise caused by mislabeling when Tri-training for network intrusion detection (NID), we propose a Tri-training algorithm based on cross entropy and K-nearest neighbors (TCK) for network intrusion detection. The proposed algorithm uses cross-entropy to replace the classification error rate to better identify the difference between the practical and predicted distributions of the model and reduce the prediction bias of mislabeled data to unlabeled data; K-nearest neighbors are used to remove the mislabeled data and reduce the number of mislabeled data. In order to verify the effectiveness of the algorithm proposed in this paper, experiments were conducted on 12 UCI datasets and NSL-KDD network intrusion datasets, and four indexes including accuracy, recall, F-measure and precision were used for comparison. The experimental results revealed that the TCK has superior performance than the conventional Tri-training algorithms and the Tri-training algorithms using only cross-entropy or K-nearest neighbor strategy.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권11호
• /
• pp.4028-4042
• /
• 2021

Aiming at the problem of software defect prediction difficulty caused by insufficient software defect marker samples and unbalanced classification, a semi-supervised software defect prediction model based on a tri-training algorithm was proposed by combining feature normalization, over-sampling technology, and a Tri-training algorithm. First, the feature normalization method is used to smooth the feature data to eliminate the influence of too large or too small feature values on the model's classification performance. Secondly, the oversampling method is used to expand and sample the data, which solves the unbalanced classification of labelled samples. Finally, the Tri-training algorithm performs machine learning on the training samples and establishes a defect prediction model. The novelty of this model is that it can effectively combine feature normalization, oversampling techniques, and the Tri-training algorithm to solve both the under-labelled sample and class imbalance problems. Simulation experiments using the NASA software defect prediction dataset show that the proposed method outperforms four existing supervised and semi-supervised learning in terms of Precision, Recall, and F-Measure values.

• 스마트미디어저널
• /
• 제10권1호
• /
• pp.16-24
• /
• 2021

word2vec 등 기존의 단어 임베딩 기법은 원시 말뭉치에 출현한 단어들만을 대상으로 각 단어를 다차원 실수 벡터 공간에 고정된 길이의 벡터로 표현하기 때문에 형태론적으로 풍부한 표현체계를 가진 언어에 대한 단어 임베딩 기법에서는 말뭉치에 출현하지 않은 단어들에 대한 단어 벡터를 표현할 때 OOV(out-of-vocabulary) 문제가 빈번하게 발생한다. 문장을 구성하는 단어 벡터들로부터 문장 벡터를 구성하는 문장 임베딩의 경우에도 OOV 단어가 포함되었을 때 문장 벡터를 정교하게 구성하지 못하는 문제점이 있다. 특히, 교착어인 한국어는 어휘형태소와 문법형태소가 결합되는 형태론적 특성 때문에 미등록어의 임베딩 기법은 성능 향상의 중요한 요인이다. 본 연구에서는 단어의 형태학적인 정보를 이용하는 방식을 문장 수준으로 확장하고 OOV 단어 문제에 강건한 병렬 Tri-LSTM 문장 임베딩을 제안한다. 한국어 감정 분석 말뭉치에 대해 성능 평가를 수행한 결과 한국어 문장 임베딩을 위한 임베딩 단위는 형태소 단위보다 문자 단위가 우수한 성능을 보였으며, 병렬 양방향 Tri-LSTM 문장 인코더는 86.17%의 감정 분석 정확도를 달성하였다.

• 한국전문물리치료학회지
• /
• 제17권4호
• /
• pp.77-87
• /
• 2010

The present case study highlights the effects of a novel Comprehensive Hand Repetitive Intensive Strengthening Training (CHRIST) on morphological changes and associated upper extremity (UE) muscle strength and motor performance in a child with spastic quadriplegic cerebral palsy (CP). The Child, a 10-year-old girl with spastic quadriplegic CP, was treated with CHRIST for 60 minutes a day, five times a week, for 5 weeks. The CHRIST was designed to improve motor function and strength. Clinical tests including the modified Wolf Test, Jebsen-Taylor Hand Function Test, and Pediatric Motor Activity Log questionnaire were used to determine motor function. Ultrasound imaging was performed to determine the changes in the cross-section area (CSA) of the extensor carpi radialis (ECR) and triceps brachii (TRI). Muscle strength was measured with a dynamometer at pretest, and post-test, and 3-month follow-up. Ultrasound imaging data showed that the CSAs of both ECR and TRI muscles were enhanced as a function of the intervention. These changes were associated with muscle strength and motor performance and their effects remained even at a 3-month follow-up test. Our results suggest that the CHRIST was effective at treating muscle atrophy, weakness and motor dysfunction in a child with spastic quadriplegic CP.

• 말소리와 음성과학
• /
• 제7권4호
• /
• pp.3-8
• /
• 2015

In this paper, we investigate an input dimension reduction method using continuous word vector in deep neural network language model. In the proposed method, continuous word vectors were generated by using Google's Word2Vec from a large training corpus to satisfy distributional hypothesis. 1-of-${\left|V\right|}$ coding discrete word vectors were replaced with their corresponding continuous word vectors. In our implementation, the input dimension was successfully reduced from 20,000 to 600 when a tri-gram language model is used with a vocabulary of 20,000 words. The total amount of time in training was reduced from 30 days to 14 days for Wall Street Journal training corpus (corpus length: 37M words).

• Bulletin of the Korean Chemical Society
• /
• 제26권4호
• /
• pp.594-598
• /
• 2005

Treatment of a solution of $CuCl_2$ in dimethyl phosphate (DMP) with DMSO under nitrogen atmosphere afforded to a light blue fluorescence powder. Slow evaporation of $H_2O$-DMSO solution of this powder resulted in blue-sky crystals of a new polymeric Cu(II) complex, with a unit cell composed of $Cu_2(DMP)_4$(DMSO), (1). The crystal and molecular structure of the complex acquired crystallographically. Compound (1) crystallizes in the monoclinic space group $P2_1$/n with a = 12.8920(11) $\AA$, b = 13.1966(11) $\AA$, c = 14.7926(13) $\AA$, $\alpha$ = 90$^{\circ}$, $\beta$ = 98.943(2)$^{\circ}$, $\gamma$ = 90$^{\circ}$, V= 2486.1(4) ${\AA}^3$, and Z = 4. A square pyramidal environment for the metal center was established by coordination of oxygen atoms of four bridging DMP ligands in the basal positions and binding a tri-centered oxygen atom of DMSO in the apical disposition of Cu(II). The sixth position was also affected by a weak interaction with the sulfur atom of another DMSO. The phosphorous atom in the bridging DMP was arranged in a deformed tetrahedron with (gg) conformation for methyl esters with $C_{2v}$ symmetry.

• 한국정보통신학회:학술대회논문집
• /
• 한국해양정보통신학회 2008년도 춘계종합학술대회 A
• /
• pp.229-232
• /
• 2008

A Neural network classification of human activity data is presented. The data acquisition system involves a tri-axial accelerometer in wireless sensor network environment. The wireless ad-hoc system has the advantage of small size, convenience for wearability and cost effectiveness. The system can further improve the range of user mobility with the inclusion of ad-hoc environment. The classification is based on the frequencies of the involved activities. The most significant Fast Fourier coefficients, of the acceleration of the body movement, are used for classification of the daily activities like, Rest walk and Run. A supervised learning approach is used. The work presents classification accuracy with the available fast batch training algorithms i.e. Levenberg-Marquardt and Resilient back propagation scheme is used for training and calculation of accuracy.

• 산업기술연구
• /
• 제35권
• /
• pp.55-58
• /
• 2015

Natural shifting of rivers has been disturbed by anthropological activities. Urbanization in Kathmandu has been encroached the natural floodplain of Bagmati and currently channeled by training walls. The study compares the change in shape of Dhobikhola, a small tributary using 1966 and 2014 satellite images. It has been found that the original shape is heavily changed over time at the beginning and end section of river under study. The river width is now fixed by training walls and roads along the banks. Using multiple data sets like satellite images and GIS analysis, these shifts can be easily detected to plan for management and restoration of physical and ecological behaviors of rivers.

• The Journal of Asian Finance, Economics and Business
• /
• 제7권6호
• /
• pp.551-560
• /
• 2020

This research investigates the impact of vocational training on wages of ethnic minority labors in emerging countries; Vietnam is the case study. The study uses secondary data from 2014 to 2018 collected through Vietnam Household Living Standards Surveys (VHLSS) conducted by the General Statistics Office. In order to analyze the impact of vocational training on wages of ethnic minority areas in Vietnam, this research creates ethnic area variables. According to Vietnamese regulations, ethnic areas are communes of 51 different provinces, inhabited by ethnic minority people. The statistics from VHLSS in 2018, show that the proportion of labors of working age with a certificate was 22.5%. The research employs Heckman Sample Selection Model to estimate the impact of vocation training on wage of labors in ethnic minority areas. The results show that vocational training plays a crucial role in improving the wages of ethnic minorities and has a positive impact. However, apart from the achieved outcomes, vocational training and job creation for ethnic minorities are not without limitations and shortcomings. Based on the findings, some recommendations to ethnic minority labors, enterprises and the Government are proposed to encourage participation in vocational training for the purpose of promoting the efficiency of the labor market.

• 정보처리학회논문지:컴퓨터 및 통신 시스템
• /
• 제12권5호
• /
• pp.165-170
• /
• 2023

최근 블록체인 기술이 부상하면서 이를 이용한 암호화폐 플랫폼이 늘어나며 화폐 거래가 활발이 이뤄지고 있다. 그러나 암호화폐의 특성을 악용한 범죄 또한 늘어나 문제가 되고 있다. 특히 피싱 스캠은 이더리움 사이버 범죄의 과반수 이상을 차지하며 주요 보안 위협원으로 여겨지고 있다. 따라서 효과적인 피싱 스캠 탐지 방법이 시급하다. 그러나 전체 이더리움 참여 계정 주소에서 라벨링된 피싱 주소의 부족으로 인한 데이터 불균형 문제로 지도학습에 충분한 데이터 제공이 어려운 상황이다. 이를 해결하기 위하여 본 논문에서는 이더리움 트랜잭션 네트워크를 고려한 효과적인 그래프 임베딩 기법인 trans2vec과 준지도 학습 모델 tri-training을 함께 사용하여 라벨링된 데이터 뿐만 아니라 라벨링되지 않은 데이터도 최대한 활용하는 피싱 스캠 탐지 방법을 제안한다.