• 한국경영과학회지
• /
• 제30권3호
• /
• pp.137-149
• /
• 2005

This study proposes a framework enhancing the accuracy of estimation for project duration by combining linear Bayesian updating scheme with the learning curve effect. Activities in a particular project might share resources in various forms and might be affected by risk factors such as weather Statistical dependence stemming from such resource or risk sharing might help us learn about the duration of upcoming activities in the Bayesian model. We illustrate, using a Monte Carlo simulation, that for partially repetitive projects a higher degree of statistical dependence among activity duration results in more variation in estimating the project duration in total, although more accurate forecasting Is achievable for the duration of an individual activity.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2004년도 ICCAS
• /
• pp.631-634
• /
• 2004

In this paper, the Bayesian recurrent neural network (BRNN) is proposed to predict time series data. Among the various traditional prediction methodologies, a neural network method is considered to be more effective in case of non-linear and non-stationary time series data. A neural network predictor requests proper learning strategy to adjust the network weights, and one need to prepare for non-linear and non-stationary evolution of network weights. The Bayesian neural network in this paper estimates not the single set of weights but the probability distributions of weights. In other words, we sets the weight vector as a state vector of state space method, and estimates its probability distributions in accordance with the Bayesian inference. This approach makes it possible to obtain more exact estimation of the weights. Moreover, in the aspect of network architecture, it is known that the recurrent feedback structure is superior to the feedforward structure for the problem of time series prediction. Therefore, the recurrent network with Bayesian inference, what we call BRNN, is expected to show higher performance than the normal neural network. To verify the performance of the proposed method, the time series data are numerically generated and a neural network predictor is applied on it. As a result, BRNN is proved to show better prediction result than common feedforward Bayesian neural network.

• 제어로봇시스템학회논문지
• /
• 제10권12호
• /
• pp.1295-1304
• /
• 2004

In this paper, the Bayesian recurrent neural network is proposed to predict time series data. A neural network predictor requests proper learning strategy to adjust the network weights, and one needs to prepare for non-linear and non-stationary evolution of network weights. The Bayesian neural network in this paper estimates not the single set of weights but the probability distributions of weights. In other words, the weights vector is set as a state vector of state space method, and its probability distributions are estimated in accordance with the particle filtering process. This approach makes it possible to obtain more exact estimation of the weights. In the aspect of network architecture, it is known that the recurrent feedback structure is superior to the feedforward structure for the problem of time series prediction. Therefore, the recurrent neural network with Bayesian inference, what we call Bayesian recurrent neural network (BRNN), is expected to show higher performance than the normal neural network. To verify the proposed method, the time series data are numerically generated and various kinds of neural network predictor are applied on it in order to be compared. As a result, feedback structure and Bayesian learning are better than feedforward structure and backpropagation learning, respectively. Consequently, it is verified that the Bayesian reccurent neural network shows better a prediction result than the common Bayesian neural network.

• 대한수학회지
• /
• 제59권2호
• /
• pp.217-233
• /
• 2022

Although machine learning shows state-of-the-art performance in a variety of fields, it is short a theoretical understanding of how machine learning works. Recently, theoretical approaches are actively being studied, and there are results for one of them, margin and its distribution. In this paper, especially we focused on the role of margin in the perturbations of inputs and parameters. We show a generalization bound for two cases, a linear model for binary classification and neural networks for multi-classification, when the inputs have normal distributed random noises. The additional generalization term caused by random noises is related to margin and exponentially inversely proportional to the noise level for binary classification. And in neural networks, the additional generalization term depends on (input dimension) × (norms of input and weights). For these results, we used the PAC-Bayesian framework. This paper is considering random noises and margin together, and it will be helpful to a better understanding of model sensitivity and the construction of robust generalization.

• Smart Structures and Systems
• /
• 제29권2호
• /
• pp.321-336
• /
• 2022

Model uncertainty is a key factor that could influence the accuracy and reliability of numerical model-based analysis. It is necessary to acquire an appropriate updating approach which could search and determine the realistic model parameter values from measurements. In this paper, the Bayesian model updating theory combined with the transitional Markov chain Monte Carlo (TMCMC) method and K-means cluster analysis is utilized in the updating of the structural model parameters. Kriging and polynomial chaos expansion (PCE) are employed to generate surrogate models to reduce the computational burden in TMCMC. The selected updating approaches are applied to three structural examples with different complexity, including a two-storey frame, a ten-storey frame, and the national stadium model. These models stand for the low-dimensional linear model, the high-dimensional linear model, and the nonlinear model, respectively. The performances of updating in these three models are assessed in terms of the prediction uncertainty, numerical efforts, and prior information. This study also investigates the updating scenarios using the analytical approach and surrogate models. The uncertainty quantification in the Bayesian approach is further discussed to verify the validity and accuracy of the surrogate models. Finally, the advantages and limitations of the surrogate model-based updating approaches are discussed for different structural complexity. The possibility of utilizing the boosting algorithm as an ensemble learning method for improving the surrogate models is also presented.

• Journal of Korea Artificial Intelligence Association
• /
• 제1권1호
• /
• pp.11-16
• /
• 2023

Accurate hospital case modeling and prediction are crucial for efficient healthcare. In this study, we demonstrate the implementation of regression analysis methods in machine learning systems utilizing mathematical statics and machine learning techniques. The developed machine learning model includes Bayesian linear, artificial neural network, decision tree, decision forest, and linear regression analysis models. Through the application of these algorithms, corresponding regression models were constructed and analyzed. The results suggest the potential of leveraging machine learning systems for medical research. The experiment aimed to create an Azure Machine Learning Studio tool for the speedy evaluation of multiple regression models. The tool faciliates the comparision of 5 types of regression models in a unified experiment and presents assessment results with performance metrics. Evaluation of regression machine learning models highlighted the advantages of boosted decision tree regression, and decision forest regression in hospital case prediction. These findings could lay the groundwork for the deliberate development of new directions in medical data processing and decision making. Furthermore, potential avenues for future research may include exploring methods such as clustering, classification, and anomaly detection in healthcare systems.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제10권4호
• /
• pp.314-318
• /
• 2010

Predicting Alpha-helicies, Beta-sheets and Turns of a proteins secondary structure is a complex non-linear task that has been approached by several techniques such as Neural Networks, Genetic Algorithms, Decision Trees and other statistical or heuristic methods. This project introduces a new machine learning method by combining Bayesian Inference with offline trained Multilayered Perceptron (MLP) models as the likelihood for secondary structure prediction of proteins. With varying window sizes of neighboring amino acid information, the information is extracted and passed back and forth between the Neural Net and the Bayesian Inference process until the posterior probability of the secondary structure converges.

• 한국콘텐츠학회논문지
• /
• 제8권5호
• /
• pp.276-286
• /
• 2008

본 논문에서는 컴퓨터 대수 시스템을 기반으로 하는 웹 저작 환경과 베이지언 추론망을 적용한 학습자 진단 환경이 포함된 이공계 수학용 적응적 이러닝 시스템 개발에 대하여 소개하였다. 본 시스템을 활용하면 교수자는 컴퓨터 대수 시스템을 수식처리 엔진으로 하며 웹을 인터페이스로 하는 이공계 수학용 웹 콘텐츠를 쉽게 생성할 수 있다. 구체적으로 선형대수, 미분방정식 및 이산수학의 영역에서 콘텐츠 개발의 예를 소개하였다. 또한 학습자의 지식 영역별 수준을 조건부 확률을 이용한 통계적 추론에 의해 진단하여 그 결과에 따라 피드백을 생성하는 적응적 이러닝 웹 콘텐츠를 만들 수 있다. 본 시스템을 사용하여 개발한 이공계 수학용 웹 콘텐츠를 평가하기 위하여 그 결과물을 대학 강의에 적용하였고, 설문지 조사를 통하여 콘텐츠 사용에 대한 학습자의 반응을 평가하였다.

• 제어로봇시스템학회논문지
• /
• 제15권9호
• /
• pp.929-938
• /
• 2009

This paper presents a novel control methodology for communication network based nonlinear systems with time delay nature. We construct a nominal nonlinear control law for representing a linear model and a reset control system which is aimed for corrective control strategy to compensate system error due to uncertain time delay through wireless communication network. Next, online neural control approach is proposed for overcoming nonstationary statistical nature in the network topology. Additionally, DBN (Dynamic Bayesian Network) technique is accomplished for modeling of its dynamics in terms of casuality, which is then utilized for estimating prediction of system output. We evaluate superiority and reliability of the proposed control approach through numerical simulation example in which a nonlinear inverted pendulum model is employed as a networked control system.

• Animal Bioscience
• /
• 제37권4호
• /
• pp.622-630
• /
• 2024

Objective: Pig breeders cannot obtain phenotypic information at the time of selection for sow lifetime productivity (SLP). They would benefit from obtaining genetic information of candidate sows. Genomic data interpreted using deep learning (DL) techniques could contribute to the genetic improvement of SLP to maximize farm profitability because DL models capture nonlinear genetic effects such as dominance and epistasis more efficiently than conventional genomic prediction methods based on linear models. This study aimed to investigate the usefulness of DL for the genomic prediction of two SLP-related traits; lifetime number of litters (LNL) and lifetime pig production (LPP). Methods: Two bivariate DL models, convolutional neural network (CNN) and local convolutional neural network (LCNN), were compared with conventional bivariate linear models (i.e., genomic best linear unbiased prediction, Bayesian ridge regression, Bayes A, and Bayes B). Phenotype and pedigree data were collected from 40,011 sows that had husbandry records. Among these, 3,652 pigs were genotyped using the PorcineSNP60K BeadChip. Results: The best predictive correlation for LNL was obtained with CNN (0.28), followed by LCNN (0.26) and conventional linear models (approximately 0.21). For LPP, the best predictive correlation was also obtained with CNN (0.29), followed by LCNN (0.27) and conventional linear models (approximately 0.25). A similar trend was observed with the mean squared error of prediction for the SLP traits. Conclusion: This study provides an example of a CNN that can outperform against the linear model-based genomic prediction approaches when the nonlinear interaction components are important because LNL and LPP exhibited strong epistatic interaction components. Additionally, our results suggest that applying bivariate DL models could also contribute to the prediction accuracy by utilizing the genetic correlation between LNL and LPP.