• International Journal of Computer Science & Network Security
• /
• v.22 no.7
• /
• pp.301-307
• /
• 2022

Energy consumption has grown alongside dramatic population increases. Statistics show that buildings in particular utilize a significant amount of energy, worldwide. Because of this, building energy prediction is crucial to best optimize utilities' energy plans and also create a predictive model for consumers. To improve energy prediction performance, this paper proposes a ResNet-LSTM model that combines residual networks (ResNets) and long short-term memory (LSTM) for energy consumption prediction. ResNets are utilized to extract complex and rich features, while LSTM has the ability to learn temporal correlation; the dense layer is used as a regression to forecast energy consumption. To make our model more robust, we employed Huber loss during the optimization process. Huber loss obtains high efficiency by handling minor errors quadratically. It also takes the absolute error for large errors to increase robustness. This makes our model less sensitive to outlier data. Our proposed system was trained on historical data to forecast energy consumption for different time series. To evaluate our proposed model, we compared our model's performance with several popular machine learning and deep learning methods such as linear regression, neural networks, decision tree, and convolutional neural networks, etc. The results show that our proposed model predicted energy consumption most accurately.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.22 no.1
• /
• pp.89-97
• /
• 2011

We analyze the performance of various techniques to overcome degradation of performance of STAP caused by nonhomogeneous clutter. The performance of NHD that used to eliminate outliers from nonhomogeneous clutter is improved by using the projection statistics(PS) that is robust to multiple outliers. The method of clutter covariance matrix estimation using a median value and the conventional method are also investigated and then compared. From the simulation results of STAP, the method of clutter covariance matrix estimation using a median value shows better performance than the conventional method for the calculation of the SINR loss, and MSMI for the single target and the multiple targets regardless of the NHD methods.

• Journal of The Korean Astronomical Society
• /
• v.48 no.1
• /
• pp.83-92
• /
• 2015

Recent large scale surveys such as Sloan Digital Sky Survey have produced homogeneous samples of multiple-image gravitationally lensed quasars with well-defined selection effects. Statistical analysis on these can yield independent constraints on cosmological parameters. Here we use the image separation statistics of lensed quasars from Sloan Digital Sky Survey Quasar Lens Search (SQLS) to derive constraints on cosmological parameters. Our analysis does not require knowledge of the magnification bias, which can only be estimated from the detailed knowledge on the quasar luminosity function at all redshifts, and includes the consideration for the bias against small image separation quasars due to selection against faint lens galaxy in the follow-up observations for confirmation. We first use the mean image separation of the lensed quasars as a function of redshift to find that cosmological models with extreme curvature are inconsistent with observed lensed quasars. We then apply the maximum likelihood test to the statistical sample of 16 lensed quasars that have both measured redshift and magnitude of lens galaxy. The likelihood incorporates the probability that the observed image separation is realized given the luminosity of the lens galaxy in the same manner as Im et al. (1997). We find that the 95% confidence range for the cosmological constant (i.e., the vacuum energy density) is $0.72{\leq}{\Omega}_{\Lambda}{\leq}1.0$ for a flat universe. We also find that the equation of state parameter can be consistent with -1 as long as the matter density ${\Omega}_m{\leq}0.4$ (95% confidence range). We conclude that the image separation statistics incorporating the brightness of lens galaxies can provide robust constraints on the cosmological parameters.

• The Journal of the Acoustical Society of Korea
• /
• v.29 no.3
• /
• pp.200-208
• /
• 2010

In this work, the minimum statistics (MS) algorithm is combined with the codebook driven short-term predictor parameter estimation (CDSTP) to design a speech enhancement algorithm that is robust against various background noise environments. The MS algorithm functions well for the stationary noise but relatively not for the non-stationary noise. The CDSTP works efficiently for the non-stationary noise, but not for the noise that was not considered in the training stage. Thus, we propose to combine CDSTP and MS. Compared with the single use of MS and CDSTP, the proposed method produces better perceptual evaluation of speech quality (PESQ) score, and especially works excellent for the mixed background noise between stationary and non-stationary noises.

• The Korean Journal of Applied Statistics
• /
• v.24 no.6
• /
• pp.1181-1196
• /
• 2011

Image fusion is the process of combining information from two or more source images of a scene into a single composite image with application to many fields, such as remote sensing, computer vision, robotics, medical imaging and defense. The most common wavelet-based fusion is discrete wavelet transform fusion in which the high frequency sub-bands and low frequency sub-bands are combined on activity measures of local windows such standard deviation and mean, respectively. However, discrete wavelet transform is not translation-invariant and it often yields block artifacts in a fused image. In this paper, we propose a robust image fusion based on the stationary wavelet transform to overcome the drawback of discrete wavelet transform. We use the activity measure of interquartile range as the robust estimator of variance in high frequency sub-bands and combine the low frequency sub-band based on the interquartile range information present in the high frequency sub-bands. We evaluate our proposed method quantitatively and qualitatively for image fusion, and compare it to some existing fusion methods. Experimental results indicate that the proposed method is more effective and can provide satisfactory fusion results.

• Journal of Korean Society for Quality Management
• /
• v.39 no.3
• /
• pp.400-411
• /
• 2011

Process variables are factors in an experiment that are not mixture components but could affect the blending properties of the mixture ingredients. For example, the effectiveness of an etching solution which is measured as an etch rate is not only a function of the proportions of the three acids that are combined to form the mixture, but also depends on the temperature of the solution and the agitation rate. Efficient designs for the mixture components-process variables experiments depend on the mixture components-process variables model which is called a combined model. We often use the product model between the canonical polynomial model for the mixture and process variables model as a combined model. In this paper we propose three starting models for the mixture components-process variables experiments. One of the starting model we are considering is the model which includes product terms up to cubic order interactions between mixture effects and the linear & pure quadratic effect of the process variables from the product model. In this paper, we propose a method for finding robust designs and practical designs with respect to D-, G-, and I-optimality for the various starting combined models and then, we find practically efficient and robust designs for estimating the regression coefficients for those models. We find the prediction capability of those recommended designs in the case of three components and three process variables to be good by checking FDS(Fraction of Design Space) plots.

• The Korean Journal of Applied Statistics
• /
• v.15 no.2
• /
• pp.395-403
• /
• 2002

In the third phase of the response surface methods, the first-order model is assumed and the curvature of the response surface is checked with a fractional factorial design augmented by centre runs. We further assume that a true model is a quadratic polynomial. To choose an optimal design, Box and Draper(1959) suggested the use of an average mean squared error (AMSE), an average of MSE of y(x) over the region of interest R. The AMSE can be partitioned into the average prediction variance (APV) and average squared bias (ASB). Since AMSE is a function of design moments, region moments and a standardized vector of parameters, it is not possible to select the design that minimizes AMSE. As a practical alternative, Box and Draper(1959) proposed minimum bias design which minimize ASB and showed that factorial design points are shrunk toward the origin for a minimum bias design. In this paper we propose a robust AMSE design which maximizes the minimum efficiency of the design with respect to a standardized vector of parameters.

• Journal of the Korea Computer Graphics Society
• /
• v.24 no.3
• /
• pp.43-48
• /
• 2018

In this paper, we present a method for removing shadows from scanned documents. Compared to the existing methods such as one based on image pyramid representation or adaptive thresholding, our method produces more robust and higher quality results. The basic idea of the approach is to use the local image statistics and to separate interesting regions from the image such as the regions around letters and figures. For the separated regions, we adaptively adjust the local brightness and contrast, and apply the sigmoid function to the intensity values as well to enhance the clarity of the image. For separated the other empty regions, we apply the gradient-base image hole filling method to fill the region with smooth color change.

• The Korean Journal of Applied Statistics
• /
• v.28 no.3
• /
• pp.561-571
• /
• 2015

When we consider an optimal design to predict the response corresponding to the point outside the design region, we are extremely careful about choosing the design criteria for selecting the support points. The assumed model and its accompanying error structure should be assumed to extend beyond the design region for the selected design criteria to be valid. Thus, we modify the existing design criteria such as extrapolation-optimality to be suited to those situations. We propose some maximin approaches in this paper. Simple and quadratic regression models are tested to find the basic characteristics of such maximin approaches. Some main findings are discussed in the conclusion.

• IE interfaces
• /
• v.23 no.3
• /
• pp.257-263
• /
• 2010

Taguchi defined a good quality as 'A correspondence of product characteristic's expected value to the objective value satisfying the minimum variance condition.' For his good quality, he suggested Taguchi Method which is called Robust design which is irrelevant to the effect of these noise factors. Taguchi Method which has many success examples and which is used by many manufacturing industry. But Optimal solution of Taguchi Method is one among the experiments which is not optimal area of experiment point. On the other hand, Response Surface Method (RSM) which has advantage to find optimal solution area experiments points by approximate polynomial regression. But Optimal of RSM is depended on initial point and RSM can not use many factors because of a great many experiment. In this paper, we combine the Taguchi Method and the Response Surface Method with each advantage which is called Taguchi-RSM. Taguchi-RSM has two step, first step to find first solution by Taguchi Method, second step to find optimal solution by RSM with initial point as first step solution. We give example using catapults.