• 제어로봇시스템학회논문지
• /
• 제13권2호
• /
• pp.173-178
• /
• 2007

This paper is to establish an appropriate model for predicting the temperature decreases in the batch transferred from the refining process to the caster in steel-making companies. Mathematical modeling of the temperature decreases between the processes is difficult, since the reaction mechanism by which the temperature changes in a molten steel batch is dynamic, uncertain and complex. Three soft computing techniques are examined using the same data, namely the multiple regression, fuzzy regression, and neural net (NN) models. To compare the accuracy of these three models, a limited number of input variables are selected from those variables significantly affecting the temperature decrease. The results show that the difference in accuracy between the three models is not statistically significant. Nonetheless, the NN model is recommended because of its adaptive ability and robustness. The method presented in this paper allows the temperature decrease to be predicted without requiring any precise metallurgical knowledge.

• 한국축산식품학회:학술대회논문집
• /
• 한국축산식품학회 2004년도 제34차 추계 국제 학술대회
• /
• pp.407-411
• /
• 2004

Ohmic heating occurs when an electric current is passes through food, resulting in a temperature rise in the product due to the conversion of the electric energy into heat. The time spent in the thawing is critical for product sterility and quality. The objective of this study is to conduct numerical modelling between the effect of ohmic thawing intensity on PTT(phase transition time) at constant concentration and the effect of matrix concentrations on PTT at constant voltage condition. the stronger ohmic thawing intensity resulted in decreasing the PTT. High ohmic intensity causes short PTT. And the higher gelatin concentration, the faster increment of PTT. A numerical modeling was executed to predict the PTT influenced by the power intensity using exponential regression and the PTT influenced by gelatin concentration using logarithmic regression. Therefore, from this numerical model of gelatin matrix, it is possible to estimate exact values extensively.

• Communications for Statistical Applications and Methods
• /
• 제10권1호
• /
• pp.239-246
• /
• 2003

Statisticians, or data miners, are often requested to assess the importances of input variables in the given supervised learning model. For the purpose, one may rely on separate ad hoc measures depending on modeling types, such as linear regressions, the neural networks or trees. Consequently, the conceptual consistency in input variable importance measures is lacking, so that the measures cannot be directly used in comparing different types of models, which is often done in data mining processes, In this short communication, we propose a unified approach to the importance measurement of input variables. Our method uses sensitivity analysis which begins by perturbing the values of input variables and monitors the output change. Research scope is limited to the models for continuous output, although it is not difficult to extend the method to supervised learning models for categorical outcomes.

• 한국정보과학회 언어공학연구회:학술대회논문집(한글 및 한국어 정보처리)
• /
• 한국정보과학회언어공학연구회 1999년도 제11회 한글 및 한국어 정보처리 학술대회
• /
• pp.191-195
• /
• 1999

자연스럽고 명료한 한국어 Text-to-Speech 변환 시스템을 위해서 음소의 지속 시간을 제어하는 일은 매우 중요하다. 음소의 지속 시간은 여러 가지 문맥 정보에 의해서 변화하므로 제어 규칙에 의존하기 보다 방대한 데이터베이스를 이용하여 통계적인 기법으로 음소의 지속 시간에 변화를 주는 요인을 찾아내려고 하는 것이 지금의 추세이다. 본 연구에서도 트리기반 모델링 방법중의 하나인 CART(classification and regression tree) 방법을 사용하여 회귀 트리를 생성하고, 생성된 트리에 기반하여 음소의 지속 시간 예측 모델과, 자연스러운 끊어 읽기를 위한 휴지 기간 예측 모델을 제안하고 있다. 실험에 사용한 음성코퍼스는 550개의 문장으로 구성되어 있으며, 이 중 428개 문장으로 회귀 트리를 학습시켰고, 나머지 122개의 문장으로 실험하였다. 모델의 평가를 위해서 실제값과 예측값과의 상관관계를 구하였더니 음소의 지속 시간을 예측하는 회귀 트리에서는 상관계수가 0.84로 계산되었고, 끊어 읽는 경계에서의 휴지 기간을 예측하는 회귀 트리에서는 상관계수가 0.63으로 나타났다.

• 한국물환경학회지
• /
• 제22권1호
• /
• pp.166-171
• /
• 2006

The objectives of the present paper were to investigate the concentration of total coliform in wastewater effluents and the effect of water chemical and physical characters in it. The most correlated parameter with total coliform was COD. It means that the wastewater treatment efficient effects on total coliform concentration. And we developed predictive model for the total coliform concentration. The estimated parameters for model were COD, temperature, nitrite, chloride, Mn and regression model equation was determined; log (Total Coli.) = 1.861+0.065[COD]+0.038[temperature]-0.0004[$Cl^-$]+3.697[Mn]-0.32 [$NO_2-N$] The developed model provided very strong correlation ($R^2:0.82$) between total coliform and regression equation. The parameters having high sensitivity were COD and temperature. So the study indicated that if the temperature and COD of wastewater effluent were known, we would estimate the concentration of total coliform and decide the most effective usage of chlorine.

• 말소리와 음성과학
• /
• 제6권2호
• /
• pp.9-20
• /
• 2014

In this paper, we propose an automatic fluency evaluation algorithm for English speaking tests. In the proposed algorithm, acoustic features are extracted from an input spoken utterance and then fluency score is computed by using support vector regression (SVR). We estimate the parameters of feature modeling and SVR using the speech signals and the corresponding scores by human raters. From the correlation analysis results, it is shown that speech rate, articulation rate, and mean length of runs are best for fluency evaluation. Experimental results show that the correlation between the human score and the SVR score is 0.87 for 3 speaking tests, which suggests the possibility of the proposed algorithm as a secondary fluency evaluation tool.

• 한국안전학회지
• /
• 제26권6호
• /
• pp.118-123
• /
• 2011

This study deals with the traffic accidents of the arterial link sections. The purpose is to comparatively analyze the characteristics and models by accident type using the data of 24 arterial links in Cheongju. In pursuing the above, this study gives particular emphasis to modeling such the accidents as the side-right-angle collision, rear-end collision and side-swipe collision. The main results are the followings. First, six accident models are developed, which are all analyzed to be statistically significant. Second, the models are comparatively evaluated using the common and specific variables by accident type.

• 한국항해항만학회지
• /
• 제33권2호
• /
• pp.153-159
• /
• 2009

The awareness of the high-value industry for container terminal leads competitiveness of container terminals to keep high fiercely. In regards to competitive factors of container terminal, the most important point among several factors is seemed to be the speed of container loading and unloading on quayside. In container terminals in Korea, the productivity shows big difference even though its condition is similar to each terminal. The objective of this paper is to find the critical factors of container terminal productivity, which is dependant upon the capability, quantity of quay crane, transfer vehicle, and so on. For this purpose, we have researched related literatures, and collected data about container terminals in South Korea. Furthermore, we tested sensitive analysis to evaluate the extent of productivity by changing independent variable. And then we established the regression model to evaluate which factor has had the biggest impact on productivity. The results of this paper can give terminal operators guideline to improve productivity.

• Structural Engineering and Mechanics
• /
• 제53권2호
• /
• pp.373-392
• /
• 2015

In the present work, structural joints have been modeled as a pair of translational and rotational springs and frequency equation of the overall system has been developed using sub-structure synthesis. It is shown that using first few natural frequencies of the system, one can obtain a set of over-determined system of equations involving the unknown stiffness parameters. Method of multi-linear regression is then applied to obtain the best estimate of the unknown stiffness parameters. The estimation procedure has been developed first for a two parameter joint model and then for a three parameter model, in which cross coupling terms are also included. Two cases of structural connections have been considered, first with a cantilever beam with support flexibility and then a pair of beams connected through lap joint. The validity of the proposed method is demonstrated through numerical simulation and by experimentation.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2007년도 하계종합학술대회 논문집
• /
• pp.457-458
• /
• 2007

This paper propose a method for representing distributed data of sensor networks. The proposed method is based on a general distributed regression framework that models sensor data by fitting a global function to each of the local measurements and explores the possible extensions of distribution regression by using complex signal representations. In order to reduce the amount of processed data and the required communication, the signal model has to be as compact as possible, with the ability to restore the arbitrary measurements. To achieve this requirement, data compression step is included, where the basis function set is changed to an overcomplete set. This have better advantages in case of nonstationary signal modeling than complete base representation.