• The Korean Journal of Applied Statistics
• /
• v.26 no.1
• /
• pp.59-70
• /
• 2013

Exit polls failed to estimate the number of seats in the National Assembly for each party in the 2012 General Election, even though they estimated it in interval. Three major broadcast companies jointly carried out exit polls, but made projections independently. The exact methods of projection were not publicly released. This paper proposes confidence intervals for the number of seats in local constituencies using the results of exit polls, and conducted simulation studies to assess the performance of the cofidence intervals. The proposed confidence intervals were applied to the real data of 2012 General Election.

• Journal of Radiation Protection and Research
• /
• v.48 no.1
• /
• pp.9-14
• /
• 2023

Background: The detectors of both computed radiography (CR) and direct digital radiography (DR) have a wide dynamic range that could tolerate high values of exposure factors without an adverse effect on image quality. Therefore, this study aims to assess patient radiation dose and proposes institutional diagnostic reference levels (DRLs) for two teaching hospitals in Ghana. Materials and Methods: CR and DR systems were utilized in this study from two teaching hospitals. The CR system was manufactured by Philips Medical Systems DMC GmbH, while the DR system was manufactured by General Electric. The entrance skin doses (ESDs) were calculated using the standard equation and the tube output measurements. Free-in-air kerma (µGy) was measured using a calibrated radiation dosimeter. The proposed institutional DRLs were estimated using 75^th percentiles values of the estimated ESDs for nine radiographic projections. Results and Discussion: The calculated DRLs were 0.4, 1.6, 3.4, 0.5, 0.4, 1.1, 1.0, 1.2, and 1.7 mGy for chest posteroanterior (PA), lumbar spine anteroposterior (AP), lumbar spine lateral (LAT), cervical spine AP, cervical spine LAT, skull PA, pelvis AP, and abdomen AP, respectively in CR system. In the DR system, the values were 0.3, 1.6, 3.1, 0.4, 0.3, 0.7, 0.6, 0.9, and 1.3 for chest PA, lumbar spine AP, lumbar spine LAT, cervical spine AP, cervical spine LAT, skull PA, pelvis AP, and abdomen AP, respectively. Conclusion: Institutional DRLs in nine radiographic projections have been proposed for two teaching hospitals in Ghana for the first time. The proposed DRLs will serve as baseline data for establishing local DRLs in the hospitals and will be a valuable tool in optimizing patient doses.

• Journal of Veterinary Clinics
• /
• v.33 no.3
• /
• pp.183-186
• /
• 2016

An 8-year-old, female lionhead rabbit with clinical sign of hematuria and vaginal discharge with/without blood was submitted to a local animal hospital. On exploratory laparotomy, three round to oval masses were observed in both uterine horns. The lumen of uterus was severely obstructed and distorted because of massive neoplastic proliferation. Histopathologically, the uterine masses revealed papillary projections along with irregular glandular structures into the lumen. The neoplastic foci were composed of numerous irregular sized neoplastic glands originated from uterine glands. These neoplastic cells showed very strong invasive tendency to muscle layer, therefore emboli of neoplastic cells were located in lymphatics. According to immunohistochemistry, the tumor cells in uterine masses demonstrated strong positive signals for cytokeratin, but negative for vimentin. Based on the gross, histopathologic and immunohistochemical features, this case was diagnosed as uterine adenocarcinoma in lionhead rabbit.

• Proceedings of the IEEK Conference
• /
• 2000.07a
• /
• pp.46-49
• /
• 2000

An iterative deblocking algorithm for DCT-compressed images using two-dimensional FIR filters adapted for local directionality of each block, is proposed. First, we introduce a set of simple lowpass filters, which are adapted for edges of different angles. In conventional deblocking methods based on lowpass-filtering and convex projections, a single filter is applied to a whole image. In the proposed method, on the other hand, a suitable filter is chosen out of the directional filters designed previously in every subimage (typically $8{\times}8$ block). Experimental results indicate that adaptive filtering improves PSNR at each iteration.

• Journal of the Korean Geographical Society
• /
• v.46 no.2
• /
• pp.212-232
• /
• 2011

The main objective of this study is to establish a population projection method based on the biregional cohort-component method and to apply it to population projections for Busan. Some drawbacks of using the net migration cohort-component method in a regional or local level population projection are demonstrated. A biregional cohort-component method, a variant of the interregional cohort-component method, is established as an alternative where in-migration and out-migration are separately considered and then are combined to produce a projection for the migration component. Predicated on the established method, population projections for Busan are undertaken for the period of 2005~2030 under three different scenarios. Considerably different projection results are obtained between the net migration and biregional methods; among others, the trend of population decline is more severe in the former than in the latter. An investigation of the temporal trend of the projected population shows that the proposed method is highly reasonable. In conclusion, the proposed method based on the biregional cohort-component method seems not only to be theoretically more robust than the net migration cohort-component method but also to be very effective in the real world application.

• Journal of Korea Water Resources Association
• /
• v.47 no.9
• /
• pp.753-766
• /
• 2014

The future mean sea level rise (MSLR) due to climate change in major harbors of Korean Peninsula has been estimated by some statistical methods in this article. Firstly, Mann-Kendall non-parametric trend test to find some trend in the observed long-term tidal data has been performed and also Bayesian change point analysis has been used also to detect the location of change points and their magnitude quantitatively. Especially, in this study, the results from Bayesian change point analysis have been applied to combine 4 future MSLR scenario projections with local MSLR data at 5 tidal gauges. This proposed procedure including Bayesian change point analysis results can improve the step for the determination of starting years of future MLSR scenario projections with 18.6-year lunar node tidal cycle and effectively consider local characteristics at each gauge. The final results by the proposed procedure in this study have shown that the future MSLR in Jeju region (Jeju tidal gauge) is in the largest increment and also the future MSLRs in Western region (Boryeong tidal gauge) and Southern region (Busan tidal gauge) are in the second largest one. Finally, it has been shown that the future MSLRs in Southern region (Yeosu tidal gauge) and Eastern region (Sokcho tidal gauge) seem to be in the relatively smallest growth among 5 gauges.

• The Journal of the Acoustical Society of Korea
• /
• v.27 no.2E
• /
• pp.35-44
• /
• 2008

The beat frequency is very important for the sound of Buddhist temple bells, but its concrete origins have not yet been fully clarified. In our research, we considered the beat of the bell at Hojobo Temple (Kanagawa Prefecture). Although its beat frequency has already been measured as 1.6 Hz, no satisfactory explanation has been offered for the factor that determines this value. In our previous research, we investigated the "Doza," the point where the bell is struck, and the "Obi," the vertical stripes around the bell, both of which are circumferential asymmetrical factors that can be visually recognized. Our investigations were carried out by using the Finite Element Method. These factors, however, could not sufficiently explain the beat frequency. Here, we first investigate the "Nyu," the many small projections on the bell, and the deviation between the centers of the inner and outer diameters of the bell. These two factors, however, were also found tobe insufficient explanations of the beat frequency. Through subsequent investigation, however, we finally clarified that the beat frequency's origin lies in the local dimension variation in the "Komazume," which is the bottom part of the bell as well as its thickest part.

• Industrial Engineering and Management Systems
• /
• v.13 no.4
• /
• pp.454-462
• /
• 2014

Retention of possible churning customer is one of the most important issues in customer relationship management, so companies try to predict churn customers using their large-scale high-dimensional data. This study focuses on dealing with large data sets by reducing the dimensionality. By using six different dimension reduction methods-Principal Component Analysis (PCA), factor analysis (FA), locally linear embedding (LLE), local tangent space alignment (LTSA), locally preserving projections (LPP), and deep auto-encoder-our experiments apply each dimension reduction method to the training data, build a classification model using the mapped data and then measure the performance using hit rate to compare the dimension reduction methods. In the result, PCA shows good performance despite its simplicity, and the deep auto-encoder gives the best overall performance. These results can be explained by the characteristics of the churn prediction data that is highly correlated and overlapped over the classes. We also proposed a simple out-of-sample extension method for the nonlinear dimension reduction methods, LLE and LTSA, utilizing the characteristic of the data.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.7
• /
• pp.3594-3607
• /
• 2017

The traditional feature extraction methods such as principal component analysis (PCA) cannot obtain the local structure of the samples, and locally linear embedding (LLE) cannot obtain the global structure of the samples. However, a common drawback of existing PCA and LLE algorithm is that they cannot deal well with the sparse problem of the samples. Therefore, by integrating the globality of PCA and the locality of LLE with a sparse constraint, we developed an improved and unsupervised difference algorithm called Sparse Difference Embedding (SDE), for dimensionality reduction of high-dimensional data in small sample size problems. Significantly differing from the existing PCA and LLE algorithms, SDE seeks to find a set of perfect projections that can not only impact the locality of intraclass and maximize the globality of interclass, but can also simultaneously use the Lasso regression to obtain a sparse transformation matrix. This characteristic makes SDE more intuitive and more powerful than PCA and LLE. At last, the proposed algorithm was estimated through experiments using the Yale and AR face image databases and the USPS handwriting digital databases. The experimental results show that SDE outperforms PCA LLE and UDP attributed to its sparse discriminating characteristics, which also indicates that the SDE is an effective method for face recognition.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.46 no.3
• /
• pp.96-104
• /
• 2009

In this paper, we propose a no-reference objective Quality assessment metrics of image. The blockiness and blurring of edge areas which are sensitive to the human visual system are modeled as step functions. Blocking and blur metrics are obtained by estimating local visibility of blockiness and edge width, For the blocking metric, horizontal and vertical blocking lines are first determined by accumulating weighted differences of adjacent pixels and then the local visibility of blockiness at the intersection of blocking lines is obtained from the total difference of amplitudes of the 2-D step function which is modelled as a blocking region. The blurred input image is first re-blurred by a Gaussian blur kernel and an edge mask image is generated. In edge blocks, the local edge width is calculated from four directional projections (horizontal, vertical and two diagonal directions) using local extrema positions. In addition, the kurtosis and SSIM are used to compute the blur metric. The final no-reference objective metric is computed after those values are combined using an appropriate function. Experimental results show that the proposed objective metrics are highly correlated to the subjective data.