• Journal of KIBIM
• /
• v.12 no.2
• /
• pp.40-48
• /
• 2022

It is very important to accurately predict construction costs in the early stages of the construction project. However, it is difficult to accurately predict construction costs with limited information from the initial stage. In recent years, with the development of machine learning technology, it has become possible to predict construction costs more accurately than before only with schematic construction characteristics. Based on machine learning technology, this study aims to analyze plans to more accurately predict construction costs by using only the factors influencing construction costs. To the end of this study, the effect of the error rate according to the activation function and the node configuration of the hidden layer was analyzed.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2009.11a
• /
• pp.445-446
• /
• 2009

IEEE 802.11 환경은 RTS/CTS(Request To Send/Clear To Send)을 지원한다. TS/CTS의 사용하면 Hidden Node Problem을 해결할 수 있지만 같은 셀안에 다른 노드를 대기상태로 만드는 False Node Problem이 발생하여 전송률을 감소시킨다. 따라서 본 논문에서는 체크포인트 방식을 사용하여 매체점유시간을 줄이는 제어프레임을 설계하고자 한다. 설계한 제어프레임의 OPNET을 사용하여 시뮬레이션하며, 기존의 제어프레임과 제안하는 제어프레임의 Delay를 비교함으로써 무선네트워크 환경에서 전송 효율을 비교 분석한다.

• 제어로봇시스템학회:학술대회논문집
• /
• 2003.10a
• /
• pp.1042-1045
• /
• 2003

In this paper, we propose the speaker identification system that uses vowel that has speaker's characteristic. System is divided to speech feature extraction part and speaker identification part. Speech feature extraction part extracts speaker's feature. Voiced speech has the characteristic that divides speakers. For vowel extraction, formants are used in voiced speech through frequency analysis. Vowel-a that different formants is extracted in text. Pitch, formant, intensity, log area ratio, LP coefficients, cepstral coefficients are used by method to draw characteristic. The cpestral coefficients that show the best performance in speaker identification among several methods are used. Speaker identification part distinguishes speaker using Neural Network. 12 order cepstral coefficients are used learning input data. Neural Network's structure is MLP and learning algorithm is BP (Backpropagation). Hidden nodes and output nodes are incremented. The nodes in the incremental learning neural network are interconnected via weighted links and each node in a layer is generally connected to each node in the succeeding layer leaving the output node to provide output for the network. Though the vowel extract and incremental learning, the proposed system uses low learning data and reduces learning time and improves identification rate.

• Journal of Intelligence and Information Systems
• /
• v.7 no.2
• /
• pp.19-33
• /
• 2001

Link weight analysis approach is suggested as a heuristic for selection of input nodes in artificial neural network for bankruptcy prediction. That is to analyze each input node\\\\`s link weight-absolute value of link weight between an input node and a hidden node in a well-trained neural network model. Prediction accuracy of three methods in this approach, -weak-linked-neurons elimination method, strong-linked-neurons selection method and integrated link weight model-is compared with that of decision tree and multivariate discrimination analysis. In result, the methods suggested in this study show higher accuracy than decision tree and multivariate discrimination analysis. Especially an integrated model has much higher accuracy than any individual models.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.05a
• /
• pp.571-573
• /
• 2018

The loss of data collected from the sensor node in the wireless sensor network environment is caused by the hidden node of the sensor node and power shortage. In order to solve these problems, researches have been actively carried out to maintain the network effectively, but there is no study on the situation where the maintenance of the network is impossible. Therefore, research is needed to infer lost data in situations where network maintenance is impossible. In this paper, use particulate matter data of specific cities to deduce lost data. Analyze the accumulated data through machine learning and identify the possibility of inferring lost data.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.34 no.3
• /
• pp.47-57
• /
• 2022

We carried out studies on prediction in concentration of dissolved oxygen (DO) with LSTM model and prediction in occurrence of hypoxia water mass (HWM) with decision tree. As results of study on prediction in DO concentration, a large number of Hidden node caused high complexity of model and required enough Epoch. And it was high accuracy in long Sequence length as prediction time step increased. The results of prediction in occurrence of HWM showed that the accuracy of nonHWM case was 66.1% in 30 day prediction, it was higher than 37.5% of HWM case. The reason is that the decision tree might overestimate DO concentration.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.8 no.3
• /
• pp.778-798
• /
• 2014

This paper proposes a novel exposed-terminal-eliminated medium access control (ETE-MAC) protocol by combining channel reservation, collision avoidance and concurrent transmissions to improve multi-access performance of the multihop wireless networks. Based on the proposed slot scheduling scheme, each node senses the control channel (CCH) or the data channel (DCH) to accurately determine whether it can send or receive the corresponding packets without collisions. Slot reservation on the CCH can be simultaneously executed with data packet transmissions on the DCH. Therefore, it resolves the hidden-terminal type and the exposed-terminal type problems efficiently, and obtains more spatial reuse of channel resources. Concurrent packet transmissions without extra network overheads are maximized. An analytical model combining Markov model and M/G/1 queuing theory is proposed to analyze its performance. The performance comparison between analysis and simulation shows that the analytical model is highly accurate. Finally, simulation results show that, the proposed protocol obviously outperforms the link-directionality-based dual-channel MAC protocol (DCP) and WiFlex in terms of the network throughput and the average packet delay.

• Proceedings of the IEEK Conference
• /
• 2001.06c
• /
• pp.25-28
• /
• 2001

RBFN has some problem that because the basis function isn't orthogonal to each others the number of used basis function goes to big. In this reason, the Wavelet Neural Network which uses the orthogonal basis function in the hidden node appears. In this paper, we propose the composition method of the actual function in hidden layer with the scaling function which can represent the region by which the several wavelet can be represented. In this method, we can decrease the size of the network with the pure several wavelet function. In addition to, when we determine the parameters of the scaling function we can process rough approximation and then the network becomes more stable. The other wavelets can be determined by the global solutions which is suitable for the suggested problem using the genetic algorithm and also, we use the back-propagation algorithm in the learning of the weights. In this step, we approximate the target function with fine tuning level. The complex neural network suggested In this paper is a new structure and important simultaneously in the point of handling the determination problem in the wavelet initialization.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2001.05a
• /
• pp.65-68
• /
• 2001

RBFN has some problem that because the basis function isnt orthogonal to each others the number of used basis function goes to big. In this reason, the Wavelet Neural Network which uses the orthogonal basis function in the hidden node appears. In this paper, we propose the composition method of the actual function in hidden layer with the scaling function which can represent the region by which the several wavelet can be represented. In this method, we can decrease the size of the network with the pure several wavelet function. In addition to, when we determine the parameters of the scaling function we can process rough approximation and then the network becomes more stable. The other wavelets can be determined by the global solutions which is suitable for the suggested problem using the genetic algorithm and also, we use the back-propagation algorithm in the learning of the weights. In this step, we approximate the target function with fine tuning level. The complex neural network suggested in this paper is a new structure and important simultaneously in the point of handling the determination problem in the wavelet initialization.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.10A
• /
• pp.830-838
• /
• 2011

Recently, studies for the network coding in the wireless network to achieve improvement of the network capacity are conducted. In this paper, we analysis considerations to apply RLNC in the wireless network. First of all, we verify whether the RLNC method in multicast is applied to distributed wireless network. In simulation results, the decoding failure can occur in the original manner of multicast. In RLNC which conducts encoding and decoding in X topology to gets rid of the decoding failure, the RLNC gain is insignificant. In this paper we also discuss considerations such as the hidden node problem, the occurrence of coding opportunity, and the RLNC overhead which are practical issues in the wireless network.