• 응용통계연구
• /
• 제31권4호
• /
• pp.447-461
• /
• 2018

국내 영화 산업 매출은 매년 증가하고 있다. 극장은 영화의 1차 판매 경로이며, 극장을 이용하는 관객 수는 부가판권에 영향을 준다. 따라서 극장을 이용하는 관객의 수는 영화 산업 매출에 직결되는 중요한 요소이다. 본 논문에서 특정일의 관객 수를 예측하기 위하여 다중선형회귀모형과 Bass 모형을 결합한 Hybrid 모형을 고려한다. 두 모형을 결합함으로써 회귀분석의 예측값을 Bass 모형의 예측값으로 보정하였다. 분석에는 개봉일이 모두 다른 세 영화를 이용하였다. All subset regression 방법을 이용해 모든 가능한 조합을 생성하고 5중 교차검증(5-fold cross validation)을 통해 5번 모형을 추정한다. 이 때 제곱근평균오차가 가장 작은 모형으로 예측값을 구한 뒤 Bass 모형의 예측값과 결합해 최종 예측값을 구하게 된다. 과거데이터가 존재할수록 Bass 모형의 가중치는 증가하면서 예측값에 보정효과를 준다는 것을 확인할 수 있었다.

• Membrane and Water Treatment
• /
• 제9권4호
• /
• pp.285-292
• /
• 2018

In the following study, Artificial Neural Network (ANN) is used for prediction of permeate flux decline during oily wastewater treatment by hybrid powdered activated carbon-microfiltration (PAC-MF) process using mullite and mullite-alumina ceramic membranes. Permeate flux is predicted as a function of time and PAC concentration. To optimize the networks performance, different transfer functions and different initial weights and biases have been tested. Totally, more than 850,000 different networks are tested for both membranes. The results showed that 10:6 and 9:20 neural networks work best for mullite and mullite-alumina ceramic membranes in PAC-MF process, respectively. These networks provide low mean squared error and high linearity between target and predicted data (high $R^2$ value). Finally, the results present that ANN provide best results ($R^2$ value equal to 0.99999) for prediction of permeation flux decline during oily wastewater treatment in PAC-MF process by ceramic membranes.

• Computers and Concrete
• /
• 제25권4호
• /
• pp.303-310
• /
• 2020

In this research, a hybrid mathematical model is derived using the higher-order polynomial kinematic model in association with soft computing technique for the prediction of best fiber volume fractions and the minimal mass of the layered composite structure. The optimal values are predicted further by taking the frequency parameter as the constraint and the projected values utilized for the computation of the eigenvalue and deflections. The optimal mass of the total layered composite and the corresponding optimal volume fractions are evaluated using the particle swarm optimization by constraining the arbitrary frequency value as mass/volume minimization functions. The degree of accuracy of the optimal model has been proven through the comparison study with published well-known research data. Further, the predicted values of volume fractions are incurred for the evaluation of the eigenvalue and the deflection data of the composite structure. To obtain the structural responses i.e. vibrational frequency and the central deflections the proposed higher-order polynomial FE model adopted. Finally, a series of numerical experimentations are carried out using the optimal fibre volume fraction for the prediction of the optimal frequencies and deflections including associated structural parameter.

• 한국음향학회지
• /
• 제20권8호
• /
• pp.12-18
• /
• 2001

본문에서는 예측형 회귀신경망과 HMM (Hidden Markov Model)의 하이브리드 네트워크인 회귀신경망 예측 HMM을 구성하였다. 회귀신경망 예측 HMM은 예측형 회귀신경망을 HMM의 각 상태마다 예측기로 정의하여 일정치인 평균벡터 대신에 과거의 특징벡터의 영향을 받아 동적으로 변화하는 신경 망에 의한 예측치를 이용하므로 학습패턴 설정자체가 시변성을 반영하는 동적 네트워크의 특성을 가진다. 따라서 음성과 같은 시계열 패턴의 인식에 유리하다. 회귀신경망 예측 HMM은 예측형 회귀신경망의 구조에 따라 Elman망 예측 HMM과 Jordan망 예측 HMM으로 구분하였다. 실험에서는 회귀신경망 예측 HMM의 상태수를 4, 5, 6으로 증가시켜 각 상태 수별로 예측차수 및 중간층 유니트 수의 변화에 따른 인식성능을 조사하였다. 실험결과 평가용 데이터에 대하여 Elman망 예측 HMM은 상태수가 6이고, 예측차수가 3차, 중간층 유니트의 수가 15차원일 때, Jordan망 예측 HMM의 경우 상태수가 5이고, 예측차수가 3차, 중간층 유니트의 수가 10차원일 때 각각 98.5%로 우수한 결과를 얻었다.

• 한국컴퓨터정보학회논문지
• /
• 제19권5호
• /
• pp.61-69
• /
• 2014

협업 필터링 추천은 사용자의 아이템에 대한 선호도를 기반으로 유사 아이템 집합 또는 유사 사용자 집합을 생성하고 이를 이용해 사용자의 특정 아이템에 대한 선호도를 예측한다. 따라서 선호도 행렬이 희박할 경우, 추천의 신뢰도는 급격히 낮아진다. 본 논문에서는 위 문제를 해결하기 위해 데이터 신뢰도 기반 가중치를 이용한 하이브리드 선호도 예측 기법을 제안한다. 선호도 예측은 유사 아이템 집합과 유사 사용자 집합을 모두 생성하고 각 집합을 통해 사용자의 선호도를 예측하며, 모델의 상황을 반영한 가중치를 이용해 각 예측치를 병합하여 수행된다. 이 기법은 사용자 선호도 예측 정확도를 높이며 선호도 행렬 희박도가 높은 상황에도 추천 서비스의 신뢰도를 유지할 수 있도록 한다. 이 기법을 바탕으로 추천 시스템을 구현하고 절대평균오차를 기준으로 서비스 신뢰도 향상을 측정하였다. 실험에서 본 기법은 Hao Ji가 제안한 기존의 기법에 비해 선호도 행렬 희박도가 84% 이상인 상황에서 평균 21.7%의 성능 향상을 보여 효과적으로 행렬 희박도 문제를 해소할 수 있음을 검증하였다.

• 대한기계학회:학술대회논문집
• /
• 대한기계학회 2007년도 춘계학술대회B
• /
• pp.2303-2308
• /
• 2007

Experiments were performed to determine the thermal (or turbulent) diffusion coefficient (TDC) and to investigate the critical heat flux (CHF) performance in the 5${\times}$5 rod bundle with 5 unheated rods which are supported by Hybrid Mixing Vane. In this study, HFC-134a fluid was used as working fluid and the fluid temperature were measured in the important subchannels. To determine the TDC value, the measured fluid temperatures were compared with the predicted values obtained from the MATRA code. The best optimized value of ${\beta}$ was found to be 0.02 by considering prediction statistics, i.e., average and standard deviations of the differences between the experimental results and code calculations. Using the best optimized value of ${\beta}$ as 0.02, the MATRA code predicts the test results of the fluid temperature within ${\pm}$1.0 % of error. According to the experimental results on CHF of 5 non-heating guide tubes, the case with non-heating guide tube showed a little good performance in terms of CHF.

• Advances in concrete construction
• /
• 제13권1호
• /
• pp.11-23
• /
• 2022

Conventional concrete needs some improvement in the mechanical properties, which can be obtained by different admixtures. However, making concrete samples costume always time and money. In this paper, different types of hybrid algorithms are applied to develop predictive models for forecasting compressive strength (CS) of concretes containing metakaolin (MK) and fly ash (FA). In this regard, three different algorithms have been used, namely multilayer perceptron (MLP), radial basis function (RBF), and support vector machine (SVR), to predict CS of concretes by considering most influencers input variables. These algorithms integrated with the grey wolf optimization (GWO) algorithm to increase the model's accuracy in predicting (GWMLP, GWRBF, and GWSVR). The proposed MLP models were implemented and evaluated in three different layers, wherein each layer, GWO, fitted the best neuron number of the hidden layer. Correspondingly, the key parameters of the SVR model are identified using the GWO method. Also, the optimization algorithm determines the hidden neurons' number and the spread value to set the RBF structure. The results show that the developed models all provide accurate predictions of the CS of concrete incorporating MK and FA with R² larger than 0.9972 and 0.9976 in the learning and testing stage, respectively. Regarding GWMLP models, the GWMLP1 model outperforms other GWMLP networks. All in all, GWSVR has the worst performance with the lowest indices, while the highest score belongs to GWRBF.

• Computers and Concrete
• /
• 제8권5호
• /
• pp.597-616
• /
• 2011

With respect to rehabilitation, strengthening and retrofitting of existing and deteriorated columns in buildings and bridges, CFRP sheets have been found effective in enhancing the performance of existing RC columns by wrapping and bonding CFRP sheets externally around the concrete. Concrete columns and piers that are confined by both lateral steel reinforcement and CFRP are sometimes referred to as "hybrid" concrete columns. With the availability of experimental data on concrete columns confined by steel reinforcement and/or CFRP, the study presents modeling using artificial neural networks (ANNs) to predict the compressive strength of hybrid circular RC columns. The prediction of the ultimate confined compressive strength of RC columns is very important especially when this value is used in estimating the capacity of structures. The present ANN model used as parameters for the confining materials the lateral steel ratio (${\rho}_s$) and the FRP volumetric ratio (${\rho}_{FRP}$). The model gave good predictions for three types of confined columns: (a) columns confined with steel reinforcement only, (b) CFRP confined columns, and (c) hybrid columns confined by both steel and CFRP. The model may be used for predicting the compressive strength of existing circular RC columns confined with steel only that will be strengthened or retrofitted using CFRP.

• 한국지능정보시스템학회:학술대회논문집
• /
• 한국지능정보시스템학회 2002년도 추계정기학술대회
• /
• pp.491-501
• /
• 2002

Case-based reasoning is emerging as a leading methodology for the application of artificial intelligence. CBR is a reasoning methodology that exploits similar experienced solutions, in the form of past cases, to solve new problems. Hybrid model achieves some convergence of the wide proliferation of credit evaluation modeling. As a result, Hybrid model showed that proposed methodology classify more accurately than any of techniques individually do. It is confirmed that proposed methodology predicts significantly better than individual techniques and the other combining methodologies. The objective of the proposed approach is to determines a set of weighting values that can best formalize the match between the input case and the previously stored cases and integrates fuzzy sit concepts into the case indexing and retrieval process. The GA is used to search for the best set of weighting values that are able to promote the association consistency among the cases. The fitness value in this study is defined as the number of old cases whose solutions match the input cases solution. In order to obtain the fitness value, many procedures have to be executed beforehand. Also this study tries to transform financial values into category ones using fuzzy logic approach fur performance of credit evaluation. Fuzzy set theory allows numerical features to be converted into fuzzy terms to simplify the matching process, and allows greater flexibility in the retrieval of candidate cases. Our proposed model is to apply an intelligent system for bankruptcy prediction.

• Journal of Power Electronics
• /
• 제16권1호
• /
• pp.217-226
• /
• 2016

A new method is proposed based on a hidden Markov model (HMM) to estimate and analyze battery states of health. Battery system health states are defined according to the relationship between internal resistance and lifetime of cells. The source data (terminal voltages and currents) can be obtained from vehicular battery models. A characteristic value extraction method is proposed for HMM. A recognition framework and testing datasets are built to test the estimation rates of different states. Test results show that the estimation rates achieved based on this method are above 90% under single conditions. The method achieves the same results under hybrid conditions. We can also use the HMMs that correspond to hybrid conditions to estimate the states under a single condition. Therefore, this method can achieve the purpose of the study in estimating battery life states. Only voltage and current are used in this method, thereby establishing its simplicity compared with other methods. The batteries can also be tested online, and the method can be used for online prediction.