• Journal of Information Processing Systems
• /
• 제11권1호
• /
• pp.39-56
• /
• 2015

Continuous multi-interval prediction (CMIP) is used to continuously predict the trend of a data stream based on various intervals simultaneously. The continuous integrated hierarchical temporal memory (CIHTM) network performs well in CMIP. However, it is not suitable for CMIP in real-time mode, especially when the number of prediction intervals is increased. In this paper, we propose a real-time integrated hierarchical temporal memory (RIHTM) network by introducing a new type of node, which is called a Zeta1FirstSpecializedQueueNode (ZFSQNode), for the real-time continuous multi-interval prediction (RCMIP) of data streams. The ZFSQNode is constructed by using a specialized circular queue (sQUEUE) together with the modules of original hierarchical temporal memory (HTM) nodes. By using a simple structure and the easy operation characteristics of the sQUEUE, entire prediction operations are integrated in the ZFSQNode. In particular, we employed only one ZFSQNode in each level of the RIHTM network during the prediction stage to generate different intervals of prediction results. The RIHTM network efficiently reduces the response time. Our performance evaluation showed that the RIHTM was satisfied to continuously predict the trend of data streams with multi-intervals in the real-time mode.

• 한국지능시스템학회논문지
• /
• 제21권4호
• /
• pp.463-468
• /
• 2011

지금까지 컨텍스트 예측 방법들은 이산 속성 컨텍스트를 대상으로 예측을 수행한 경우와 연속 속성 컨텍스트를 대상으로 예측을 수행한 경우로 나뉘어서 발전되어 왔다. 대부분의 예측 방법들은 컨텍스트의 획득 환경이나 특성에 맞게 특정 도메인에서 각각 예측 알고리즘을 작성하여 사용하여 왔기 때문에, 다양한 환경과 특성을 갖는 사용자의 컨텍스트를 대상으로 예측을 수행하기가 어렵다. 본 논문에서는 특정 도메인이나 컨텍스트의 특성에 국한되지 않고 이산 속성이나 연속 속성 컨텍스트들에 모두 적용 가능한 컨텍스트 예측 방법을 제안한다. 이를 위해 컨텍스트 속성간의 연관규칙을 고려하여 컨텍스트를 시퀀스로 생성하고, 컨텍스트 속성별 가변 가중치를 적용시켜 시퀀스 매칭 기반의 컨텍스트 예측을 수행한다. 제안한 방법을 평가하기 위해 이산 속성 컨텍스트와 연속 속성 컨텍스트에 각각 시뮬레이션한 결과 이산 속성 컨텍스트에서 80.12%, 연속 속성 컨텍스트에서 81.43%의 예측 정확도로 기존 예측방법들과 비슷한 성능을 보였다.

• 정보처리학회논문지:컴퓨터 및 통신 시스템
• /
• 제1권1호
• /
• pp.11-20
• /
• 2012

스트림 데이터는 시간에 따라 연속적으로 변화하는 일련의 값들로 나타난다. 이러한 스트림 데이터의 특성상 다양한 시간 간격의 기준에 따라 계속적으로 그 동향이 달라질 수 있다. 이 때문에 스트림 데이터의 추세 예측은 간격이 갱신될 때 마다 연속적인 환경에서 여러 간격들을 기준으로 동시에 이루어지는 연속 다중 예측(Continuous Multiple Prediction, CMP)이 지원되어야 한다. 본 논문은 스트림 데이터의 연속 다중 예측을 효과적으로 지원하기 위하여, 신피질 학습 모델인 계층형 시간적 메모리(Hierarchical Temporal Memory, HTM) 모델을 확장하여 연속통합 HTM(Continuous Integrated HTM, CIHTM) 네트워크를 제안한다. 이를 위해 우리는 HTM 네트워크를 구성하는 기존 노드들 외에 새롭게 이동 벡터 파일 센서, 시공간 분류 노드, 다중 통합 노드를 고안하였다. 그리고 이들을 바탕으로 CIHTM 네트워크의 학습과 추론 알고리즘을 개발하였다.

• 설비공학논문집
• /
• 제13권9호
• /
• pp.836-844
• /
• 2001

For the radiant floor heating system, the possibility of suboptimal prediction control was investigated by computer simulation and experiment. For this study, TRANSYS program was used and an experimental facility consisting of two rooms (3$\times$4.4$\times$2.8m) was built. The facility enabled simultaneous comparison of two different control strategies which implemented in a separate room. Results showed that outdoor air temperature prediction control was superior to the conventional outdoor air temperature compensation control for radiant floor heating system. However, more research for fine prediction of outside air temperature was required in the future.

• Earthquakes and Structures
• /
• 제22권6호
• /
• pp.609-623
• /
• 2022

To study the empirical seismic fragility of a reinforced concrete girder bridge, based on the theory of numerical analysis and probability modelling, a regression fragility method of a rapid fragility prediction model (Gaussian first-order regression probability model) considering empirical seismic damage is proposed. A total of 1,069 reinforced concrete girder bridges of 22 highways were used to verify the model, and the vulnerability function, plane, surface and curve model of reinforced concrete girder bridges (simple supported girder bridges and continuous girder bridges) considering the number of samples in multiple intensity regions were established. The new empirical seismic damage probability matrix and curve models of observation frequency and damage exceeding probability are developed in multiple intensity regions. A comparative vulnerability analysis between simple supported girder bridges and continuous girder bridges is provided. Depending on the theory of the regional mean seismic damage index matrix model, the empirical seismic damage prediction probability matrix is embedded in the multidimensional mean seismic damage index matrix model, and the regional rapid prediction matrix and curve of reinforced concrete girder bridges, simple supported girder bridges and continuous girder bridges in multiple intensity regions based on mean seismic damage index parameters are developed. The established multidimensional group bridge vulnerability model can be used to quantify and predict the fragility of bridges in multiple intensity regions and the fragility assessment of regional group reinforced concrete girder bridges in the future.

• 한국멀티미디어학회논문지
• /
• 제14권9호
• /
• pp.1152-1164
• /
• 2011

주식 가격은 연속적으로 변화하는 스트림 데이터이다. 이러한 데이터의 특성상 시간의 흐름에 따라 주식 가격의 동향이 달라질 수 있기 때문에 주식 가격 동향의 예측은 가격이 갱신될 때 마다 연속적으로 이루어져야 한다. 본 논문은 HTM 모델을 이용하여 원하는 종목의 주식 가격 동향을 설정된 구간 간격에 따라 연속적으로 주식 가격 동향을 예측하는 새로운 방법을 제안한다. 이를 위해 먼저 정규화 과정을 거친 후 그 결과를 스트림 센서로 전달하는 선처리기와 연속적인 입력 데이터를 효과적으로 처리할 수 있는 스트림 센서를 제시한다. 또한, 각 레벨별 예측 결과를 저장하여 상위 단계로 전달하는 선 예측 저장 노드를 고안하고 이를 이용하여 주식 가격 동향을 예측하는 HTM 네트워크를 제시한다. 그리고 본 시스템을 실제 주식 가격으로 실험하여 그 성능을 제시한다.

• 대한기계학회논문집
• /
• 제19권5호
• /
• pp.1182-1189
• /
• 1995

Several methods have been developed to predict the creep rupture time of the steam pipes in thermal power plant. However, existing creep life prediction methods give very conservative value at operating stress of power plant and creep rupture strain cannot be well estimated. Therefore, in this study, creep rupture time and strain prediction method accounting for material damage and grain boundary sliding is newly proposed and compared with the existing experimental data. The creep damage evolves by continuous cavity nucleation and constrained cavity growth. The results showed good correlation between the theoretically predicted creep rupture time and the experimental data. And creep rupture strain may be well estimated by using the proposed method.

• 소음진동
• /
• 제7권6호
• /
• pp.1001-1006
• /
• 1997

Railroad noise is one of the main causes of environmental impact. Whenever a new railroad line is planned or a housing project near an existing railroad is proposed, an estimate of the relevant noise levels is usually required. For this, it is necessary to quantify those paramters that affect the railroad noise. This paper presents an accurate prediction of railroad noise.

• 한국주조공학회지
• /
• 제11권1호
• /
• pp.44-53
• /
• 1991

The proper feeding conditions for thin Al-Alloy (AA336, JIS AC8A) castings in permanent mold were investigated to eliminate microshrinkage porosity. 5mm-thick plates (200mm long, 60mm wide) were cast with increasing padding taper from 0 to 5% under different conditions : (1) constant mold temperature of $350^{\circ}C$, (2) continuous production with uniform mold thickness (10mm), (3) continuous production with a negative taper of 2.5% in mold thickness (thickness decreasing in direction to riser). The test casting were machined off to the midplane and the shrinkage porosity was examined visually. The critical padding taper which can just eliminate the shrinkage porosity was determined for each condition, i.e. : (1) 4.5% at the constant mold temperature, (2) 3.5% for continuous production with the uniform mold thickness (3) 1.5% for continuous production with the taper in mold thickness. A computer simulation by a finite difference analysis program was applied to the test casting. The liquid fraction gradient (LFG) and the temperature gradient divided by the square root of the cooling rate (G /SR) were calculated at the end of solidification and compared with the shrinkage porosity area in the castings. For the case of constant mold temperature, LFG is a better parameter to predict shrinkage porosity than G /SR and its critical value is around 11%/cm. But for the case of continuous production, neither LFG nor G /SR could be a reliable parameter. The experimental results about the critical padding taper are of practical interest for designing permanent molds and castings. The computer simulation results stimulate further research to be directed on the prediction of centerline microshrinkage porosity in continuous production.

• Nuclear Engineering and Technology
• /
• 제56권6호
• /
• pp.2343-2351
• /
• 2024

Accurately predicting the thermal hydraulic parameters of a transient reactor core under different working conditions is the first step toward reactor safety. Mass flow rate and temperature are important parameters of core thermal hydraulics, which have often been modeled as time series prediction problems. This study aims to achieve accurate and continuous prediction of core thermal hydraulic parameters under instantaneous conditions, as well as test the feasibility of a newly constructed gated recurrent unit (GRU) model based on the soft attention mechanism for core parameter predictions. Herein, the China Experimental Fast Reactor (CEFR) is used as the research object, and CEFR 1/2 core was taken as subject to carry out continuous predictive analysis of thermal parameters under transient conditions., while the subchannel analysis code named SUBCHANFLOW is used to generate the time series of core thermal-hydraulic parameters. The GRU model is used to predict the mass flow and temperature time series of the core. The results show that compared to the adaptive radial basis function neural network, the GRU network model produces better prediction results. The average relative error for temperature is less than 0.5 % when the step size is 3, and the prediction effect is better within 15 s. The average relative error of mass flow rate is less than 5 % when the step size is 10, and the prediction effect is better in the subsequent 12 s. The GRU model not only shows a higher prediction accuracy, but also captures the trends of the dynamic time series, which is useful for maintaining reactor safety and preventing nuclear power plant accidents. Furthermore, it can provide long-term continuous predictions under transient reactor conditions, which is useful for engineering applications and improving reactor safety.