• 제어로봇시스템학회:학술대회논문집
• /
• 1998.10a
• /
• pp.254-259
• /
• 1998

A new design methology is proposed to identify the structure and parameters of fuzzy model using PNN and a fuzzy inference method. The PNN is the extended structure of the GMDH(Group Method of Data Handling), and uses several types of polynomials such as linear, quadratic and cubic besides the biquadratic polynomial used in the GMDH. The FPNN(Fuzzy Polynomial Neural Networks) algorithm uses PNN(Polynomial Neural networks) structure and a fuzzy inference method. In the fuzzy inference method, the simplified and regression polynomial inference methods are used. Here a regression polynomial inference is based on consequence of fuzzy rules with a polynomial equations such as linear, quadratic and cubic equation. Each node of the FPNN is defined as fuzzy rules and its structure is a kind of neuro-fuzzy architecture. In this paper, we will consider a model that combines the advantage of both FPNN and PNN. Also we use the training and testing data set to obtain a balance between the approximation and generalization of process model. Several numerical examples are used to evaluate the performance of the our proposed model.

• Journal of Environmental Health Sciences
• /
• v.38 no.2
• /
• pp.136-141
• /
• 2012

Objectives: The objective of this study was to investigate whether the ammonia nitrogen concentration of aqueous samples such as drinking water can be determined by measuring the saturation of the samples colored by indophenol method. Methods: A color saturation measurement system was constructed by connecting a notebook computer to an image acquisition device composed of a PC camera and a light source, and was then used to measure the saturation of samples colored by blue indophenol complex. Results: Between two available light sources, a fluorescent lamp was selected due to its demonstrating better linearity between color saturation and ammonia nitrogen concentration. Prediction by quadratic regression was more accurate than by linear regression, and prediction by quadratic regression in the concentration range of 0.1-1.0 $mg/l$ was more accurate than in the concentration range of 0.0-1.0 $mg/l$. Regression-based predictions over 0.25 $mg/l$, 0.55 $mg/l$ and 0.75 $mg/l$ concentrations were implemented both by spectrophotometric method and by measuring color saturation. In the case of 0.25 $mg/l$, the predicted concentration by spectrophotometric method was $0.256{\pm}0.0076\;mg/l$ and the predicted concentration by measuring color saturation was $0.246{\pm}0.0086\;mg/l$ (p=0.051). In the case of 0.55 $mg/l$, they were $0.561{\pm}0.0068\;mg/l$ and $0.564{\pm}0.0166\;mg/l$ (p=0.660). In the case of 0.75 $mg/l$, they were $0.755{\pm}0.0139\;mg/l$ and $0.762{\pm}0.0088\;mg/l$ (p=0.215). Conclusions: There were no statistically significant differences (p>0.05) between the data from the two methods in all three of the concentrations. Therefore, the color saturation measurement method proposed in this paper may be considered applicable for determining the ammonia nitrogen concentration of aqueous samples such as drinking water.

• Proceedings of the KIEE Conference
• /
• 1998.11b
• /
• pp.675-677
• /
• 1998

In this paper, we propose the fuzzy inference algorithm with multi-layer structure. MFIS(Multi-layer Fuzzy Inference System) uses PNN(Polynomial Neural networks) structure and the fuzzy inference method. The PNN is the extended structure of the GMDH(Group Method of Data Hendling), and uses several types of polynomials such as linear, quadratic and cubic, as well as the biquadratic polynomial used in the GMDH. In the fuzzy inference method, the simplified and regression polynomial inference methods are used. Here, the regression polynomial inference is based on consequence of fuzzy rules with the polynomial equations such as linear, quadratic and cubic equation. Each node of the MFIS is defined as fuzzy rules and its structure is a kind of neuro-fuzzy structure. We use the training and testing data set to obtain a balance between the approximation and the generalization of process model. Several numerical examples are used to evaluate the performance of the our proposed model.

• The Korean Journal of Applied Statistics
• /
• v.5 no.1
• /
• pp.29-39
• /
• 1992

A testing and estimation procedure is considered for changes at unknown time point in simple linear regression model. A test statistic of quadratic form is suggested. We also discuss the asymptotic distribution and its level control. The proposed method is compared with the likelihood ratio test through a example.

• International Journal of High-Rise Buildings
• /
• v.11 no.4
• /
• pp.315-320
• /
• 2022

With the rapid development of urbanization the problems of pedestrian-level wind comfort and natural ventilation of tall buildings are becoming increasingly prominent. The velocity at the pedestrian level ($\overline{MVR}$) and variation of wind pressure coefficients $\overline{{\Delta}C_p}$ between windward and leeward surfaces of tall buildings were investigated systematically through numerical simulations. The examined parameters included building density ρ, height ratio of building α_H, width ratio of building α_B, and wind direction θ. The linear and quadratic regression analyses of $\overline{MVR}$ and $\overline{{\Delta}C_p}$ were conducted. The quadratic regression had better performance in predicting $\overline{MVR}$ and $\overline{{\Delta}C_p}$ than the linear regression. $\overline{MVR}$ and $\overline{{\Delta}C_p}$ were optimized by the NSGA-II algorithm. The LINMAP and TOPSIS decision-making methods demonstrated better capability than the Shannon's entropy approach. The final optimal design parameters of buildings were ρ = 20%, α_H = 4.5, and α_B = 1, and the wind direction was θ = 10°. The proposed method could be used for the optimization of pedestrian-level wind comfort and natural ventilation in a residential area.

• Proceedings of the KSME Conference
• /
• 2004.11a
• /
• pp.852-857
• /
• 2004

A new and efficient method for estimating the statistical moments of a system performance function has been developed. The method consists of two steps: (1) An approximate response surface is generated by a quadratic regression model, and (2) the statistical moments of the regression model are then calculated by experimental design techniques proposed by Seo and $Kwak^{(4)}$. In this approach, the size of experimental region affects the accuracy of the statistical moments. Therefore, the region size should be selected suitably. The D-optimal design and the central composite design are adopted over the selected experimental region for the regression model. Finally, the Pearson system is adopted to decide the distribution type of the system performance function and to analyze structural reliability.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.4 no.1
• /
• pp.57-66
• /
• 2001

This study used four theoretical models, such as two-point linear model, linear regression model, quadratic regression model and cubic regression model which are presented from The Ministry of Science and Technology, for extraction of urban surface temperature from Landsat TM band 6 image. Through correlation and regression analysis between result of four models and AWS(automatic weather station) observation data, this study could verify spatial distribution characteristic of urban surface temperature using GIS spatial analysis method. The result of analysis for surface temperature by landcover showed that the urban and the barren land belonged to the highest surface temperature class. And there was also -0.85 correlation in the result of correlation analysis between surface temperature and NDVI. In this result, the meteorological environmental characteristics wuld be regarded as one of the important factor in urban planning.

• Proceedings of the IEEK Conference
• /
• 2002.06c
• /
• pp.135-138
• /
• 2002

This paper addresses incremental learning method for regression. SVM(support vector machine) is a recently proposed learning method. In general training a support vector machine requires solving a QP (quadratic programing) problem. For very large dataset or incremental dataset, solving QP problems may be inconvenient. So this paper presents an incremental support vector learning method for function approximation problems.

• Asian-Australasian Journal of Animal Sciences
• /
• v.16 no.6
• /
• pp.894-902
• /
• 2003

Prebiotics (peptides, N-acetyglucoamine, fructo-oligosaccharides, isomalto-oligosaccharides and galactooligosaccharides) were added to skim milk in order to improve the growth rate of contained Lactobacillus acidophilus, Lactobacillus casei, Bifidobacterium longum and Bifidobacterium bifidum. The purpose of this research was to study the potential synergy between probiotics and prebiotics when present in milk, and to apply modern optimization techniques to obtain optimal design and performance for the growth rate of the probiotics using a response surface-modeling technique. To carry out response surface modeling, the regression method was performed on experimental results to build mathematical models. The models were then formulated as an objective function in an optimization problem that was consequently optimized using a genetic algorithm and sequential quadratic programming approach to obtain the maximum growth rate of the probiotics. The results showed that the quadratic models appeared to have the most accurate response surface fit. Both SQP and GA were able to identify the optimal combination of prebiotics to stimulate the growth of probiotics in milk. Comparing both methods, SQP appeared to be more efficient than GA at such a task.

• Journal of the Korean Data and Information Science Society
• /
• v.22 no.6
• /
• pp.1223-1232
• /
• 2011

Tong and Wang's estimator (2005) is a new approach to estimate the error variance using least squares method such that a simple linear regression is asymptotically derived from Rice's lag- estimator (1984). Their estimator highly depends on the setting of a regressor and weights in small sample sizes. In this article, we propose a new approach via a local quadratic approximation to set regressors in a small sample case. We estimate the error variance as the intercept using a ridge regression because the regressors have the problem of multicollinearity. From the small simulation study, the performance of our approach with some existing methods is better in small sample cases and comparable in large cases. More research is required on unequally spaced points.