• KIPS Transactions on Computer and Communication Systems
• /
• v.9 no.1
• /
• pp.17-24
• /
• 2020

In this paper, we present a movie rating prediction architecture based on a convolutional neural network (CNN). Our prediction architecture extends TextCNN, a popular CNN-based architecture for sentence classification, in three aspects. First, character embeddings are utilized to cover many variants of words since reviews are short and not well-written linguistically. Second, the attention mechanism (i.e., squeeze-and-excitation) is adopted to focus on important features. Third, a scoring function is proposed to convert the output of an activation function to a review score in a certain range (1-10). We evaluated our prediction architecture on a movie review dataset and achieved a low MSE (e.g., 3.3841) compared with an existing method. It showed the superiority of our movie rating prediction architecture.

• Computational Structural Engineering
• /
• v.9 no.2
• /
• pp.93-101
• /
• 1996

An artificial neural network is a computational model that mimics the biological system of the brain and it consists of a number of interconnected processing units where it can reasonably infer by them. Because the neural network is particularly useful for evaluating systems with a multitude of nonlinear variables, it can be used in experimental results predictions, in structural planning and in optimum design of structures. This paper describes the basic theory related to the neural networks and discusses the applicability of neural networks to predict the ultimate shear capacity of the precast concrete vertical joints by comparing the neural networks with a conventional method such as regression.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.12
• /
• pp.579-586
• /
• 2021

Time series forecasting refers to predicting future time information based on past time information. Accurately predicting future information is crucial because it is used for establishing strategies or making policy decisions in various fields. Recently, a transformer model has been mainly studied for a time series prediction model. However, the existing transformer model has a limitation in that it has an auto-regressive structure in which the output result is input again when the prediction sequence is output. This limitation causes a problem in that accuracy is lowered when predicting a distant time point. This paper proposes a sequential decoding model focusing on the style transformation technique to handle these problems and make more precise time series forecasting. The proposed model has a structure in which the contents of past data are extracted from the transformer-encoder and reflected in the style-based decoder to generate the predictive sequence. Unlike the decoder structure of the conventional auto-regressive transformer, this structure has the advantage of being able to more accurately predict information from a distant view because the prediction sequence is output all at once. As a result of conducting a prediction experiment with various time series datasets with different data characteristics, it was shown that the model presented in this paper has better prediction accuracy than other existing time series prediction models.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.14 no.3
• /
• pp.559-564
• /
• 2010

The GMDH(Group Method of Data Handling) algorithm can be used to predict the complex nonlinear systems. The traditional GMDH algorithm produces the prdicted output of the system model in the output layer through the input layer and the intermediate layers as the prescribed process. The outputs of each layer are produced only by the outputs of the former layer. However, in the traditional GMDH algorithm, though the optimal structure of each layer is derived, the overall structure may not be derived optimally. To overcome this problem, GMDH prediction model which has the overall optimal structure is constructed by feeding back the error between the predicted output and the real output. This can make the prediction more precise. The capability improvement of the proposed algorithm compared to the traditional algorithm is verified through computer simulation.

• Journal of the Korean Society of Propulsion Engineers
• /
• v.1 no.1
• /
• pp.111-121
• /
• 1997

An investigation was performed to study the predicting the joint strength of mechanical fasteners. Bearing failure is most important failure mode for designing joint. So in this study, the prediction method in consideration with bearing failure was chosen. In the proposed method, the characteristic length is combined with the Yamada-Sun failure criterion, Tsai-Hill failure criterion and characteristic length for Tension and Compression is determined from investigation. Especially the length of compression is determined from the "bearing failure test" that newly conceived to take bearing failure into consideration. The proposed prediction method was applied to quasi-isotropic carbon/epoxy joint showing net-tension and bearing failure experimentally. Good agreement was found between the predicted and experimental result for each joint geometry. geometry.

• Proceeding of EDISON Challenge
• /
• 2014.03a
• /
• pp.399-408
• /
• 2014

신약을 개발하거나 단백질 구조를 예측하는데 Molecular Mechanics (MM)의 방법을 사용한다. 하지만, MM 만으로는 자연현상에서 일어나는 결과를 정확하게 기술하기 어렵다. 본 연구는 기존의 MM 방법으로는 정확히 예측이 불가능한 비 공유결합 중 하나인 ${\pi}-{\pi}$ interaction을 양자역학 계산을 통해 정확한 예측이 가능한지 알아보았다. ${\pi}-{\pi}$ interaction이란 생채 내, 의약 화합물에서 발견되는 결합이기 때문에, 단백질과 결합하는 구조의 예측에 중요하다고 할 수 있다. 본 실험은 ${\pi}-{\pi}$ interaction을 갖는 Sandwich, T shape, 그리고 Parallel displaced 세 가지 모형과 각각의 모형 아래에 분자를 하나 더 쌓은 모형을 추가하여 양자역학 재산을 수행하였다. 양자역학 계산은 DFT의 세가지 함수 M06_2X, M05_2X, B3LYP를 이용하였다. 실험결과에서 세 가지 함수가 각기 다른 결과를 보였는데, 상대적으로 B3LYP의 경우에는 세가지 모델에서 모두 제대로 된 에너지 변화를 계산하지 못하였으며, M06_2X와 M05_2X의 결과에서는 거리에 따른 ${\pi}-{\pi}$ interaction 에너지의 변화를 정확하게 계산하였다. 이러한 결과를 바탕으로, 양자역학의 방법을 통해 MM에서는 예측이 불가능한 ${\pi}-{\pi}$ interaction을 계산 할 수 있고 이 부분을 고려하여 화합물 간의 결합구조를 예측을 향상시킬 수 있다.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.28 no.6
• /
• pp.645-657
• /
• 2015

In this study floor impact noise and structure acceleration response of bare concrete slabs were predicted by using Finite Element Analysis(FEA). Prediction results were compared with experimental results to prove the accuracy of numerical model. Acoustic absorption were addressed by using panel impedance coefficients with frequency characteristics and structural modal damping of numerical model were applied by modal testing results and analysis of prediction and test results. By using frequency response function, the floor acceleration and acoustic pressure responses for various impact sources were calculated at the same time. In the FEA, the natural frequencies and the shapes of vibration and acoustic modes can be estimated through the eigen-value analysis, and it can be visually seen the vibration and sound pressure field and the contribution of major modes.

• Journal of Digital Convergence
• /
• v.11 no.9
• /
• pp.103-111
• /
• 2013

As the demand-oriented management has been getting important in Supply Chain Management (SCM), various forecasting methods have been suggested including regression analyses. However, dependency structures among variables have been captured by a correlation coefficient, only. It results in inaccurate demand predictions. This paper suggests a new and effective forecasting modeling framework using student's t-copula function. In order to show overall modeling procedures framework, heavy tail typed numerical data and its copula estimations are provided. The suggested methodology can contribute to decrease the bullwhip effect and to stabilize volatile environment in a supply chain network.

• Proceedings of the IEEK Conference
• /
• 2001.09a
• /
• pp.257-260
• /
• 2001

본 논문에서는 2 채널 MAIN 프로필 MPEG-2 AAC 복호화 시스템의 구조를 제안하고 구현하였다. 복호화 알고리듬의 구조적인 모듈화에 근거하여, 시스템 설계 과정에서 전체 시스템을 3 개의 하드웨어 모듈로 분할하였다. 전체 시스템은 허프만 복호화기, 예측기, 20 비트 고정소수점 DSP 코어로 이루어져 있다. 허프만 복호화기는 주어진 작업을 1 클럭 사이클 내에 수행할 수 있는 고속의 하드와이어드 모듈이고, 예측기는 높은 해상도를 가지고 다른 모듈들과 병렬처리가 가능한 구조를 가진 모듈이다. 구현된 시스템은 16.9 MIPS 로 2 채널의 MPEG-2 AAC 비트열을 고음질로 복호화할 수 있다.

• Journal of Welding and Joining
• /
• v.6 no.1
• /
• pp.11-20
• /
• 1988

이 글에서는 Butt joint에서 발생되는 가로수축, 각변형 그리고 bowing의 관찰을 통해 mechanism을 서술하였고, box beam, damper blade, bulk head, ball tank, 그리고 cylindrical column 등 실구조물의 변형문제에 접근하여 변형예측, 측정 data제시, 그리고 그 해석을 통해 구조물의 변형 크기를 인식케 하고, 동일한 구조물의 제작에 guide가 되고자 하였다. 변형에 대해서는 일률적인 방지방법이 없기 때문에 우선은 구조물의 제작시마다 변형의 크기와 향상에 대한 예측과 함께 변형계측이 계속되어 data가 축적된다면 차후 그 이용 및 응용 효과는 지대할 것이다.