• Journal of the Semiconductor & Display Technology
• /
• v.22 no.3
• /
• pp.142-148
• /
• 2023

In this paper, we identify performance issues in executing compute kernels from PolyBench, which includes compute kernels that are the core computational units of various data-intensive workloads, such as deep learning and data-intensive applications, on Processing-in-Memory (PIM) devices. Therefore, using our in-house simulator, we measured and compared the various performance metrics of workloads based on traditional out-of-order and in-order processors with Processing-in-Memory-based systems. As a result, the PIM-based system improves performance compared to other computing models due to the short-term data reuse characteristic of computational kernels from PolyBench. However, some kernels perform poorly in PIM-based systems without a multi-layer cache hierarchy due to some kernel's long-term data reuse characteristics. Hence, our evaluation and analysis results suggest that further research should consider dynamic and workload pattern adaptive approaches to overcome performance degradation from computational kernels with long-term data reuse characteristics and hidden data locality.

• Interdisciplinary Bio Central
• /
• v.4 no.1
• /
• pp.2.1-2.7
• /
• 2012

Introduction: Histone deacetylases (HDAC) are a class of enzymes that remove acetyl groups from ${\varepsilon}$-N-acetyl lysine amino acids of histone proteins. Their action is opposite to that of histone acetyltransferase that adds acetyl groups to these lysines. Only few HDAC inhibitors are approved and used as anti-cancer therapeutics. Thus, discovery of new and potential HDAC inhibitors are necessary in the effective treatment of cancer. Materials and Methods: This study proposed a method using support vector machine (SVM) to classify HDAC8 inhibitors and non-inhibitors in early-phase virtual compound filtering and screening. The 100 experimentally known HDAC8 inhibitors including 52 inhibitors and 48 non-inhibitors were used in this study. A set of molecular descriptors was calculated for all compounds in the dataset using ADRIANA. Code of Molecular Networks. Different kernel functions available from SVM Tools of free support vector machine software and training and test sets of varying size were used in model generation and validation. Results and Conclusion: The best model obtained using kernel functions has shown 75% of accuracy on test set prediction. The other models have also displayed good prediction over the test set compounds. The results of this study can be used as simple and effective filters in the drug discovery process.

• Journal of the Korea Convergence Society
• /
• v.7 no.4
• /
• pp.1-7
• /
• 2016

Structural MRI(sMRI) imaging is used to extract morphometric features after Grey Matter (GM), White Matter (WM) for several univariate and multivariate method, and Cerebro-spinal Fluid (CSF) segmentation. A new approach is applied for the diagnosis of very mild to mild AD. We propose the classification method of Alzheimer disease patients from normal controls by combining morphometric features and Gaussian Mixture Models parameters along with MMSE (Mini Mental State Examination) score. The combined features are fed into Multi-kernel SVM classifier after getting rid of curse of dimensionality using principal component analysis. The experimenral results of the proposed diagnosis method yield up to 96% stratification accuracy with Multi-kernel SVM along with high sensitivity and specificity above 90%.

• The Journal of the Korea Contents Association
• /
• v.17 no.10
• /
• pp.1-8
• /
• 2017

As the recent advancement of military equipment has been accelerated, it is becoming more important to act as an inspection device that verifies the performance of equipment in real time. Most of the inspection equipments were developed on the Windows OS based system. considering development convenience and development period. However, sice the data communication between these models occurs asynchronously, there is a problem that it is difficult to guarantee real-time performance of the window-based inspection equiment. To solve these problems, we use real-time commercial solutions to guarantee the real-time performance of Windows-based inspection equipment. In this paper, we propose a method of designing and implementing the inspection equipment software based on Real-Time implanted Kernel-Multi Processor (RTiK-MP) operating in Windows environment. In addition, real-time performance data accuracy was measured through a high-speed communication tool and interlocking test to verify the performance of the inspection device based on the real-time porting kernel.

• Journal of the Optical Society of Korea
• /
• v.19 no.2
• /
• pp.136-143
• /
• 2015

Image deblurring by a deconvolution method requires accurate knowledge of the blur kernel. Existing point spread function (PSF) models in the literature corresponding to lens aberrations and defocus are either parameterized and spatially invariant or spatially varying but discretely defined. In this paper, a parameterized model is developed and presented for a PSF which is spatially varying due to lens aberrations and defocus in an imaging system. The model is established from the Seidel third-order aberration coefficient and the Hu moment. A skew normal Gauss model is selected for parameterized PSF geometry structure. The accuracy of the model is demonstrated with simulations and measurements for a defocused infrared camera and a single spherical lens digital camera. Compared with optical software Code V, the visual results of two optical systems validate our analysis and proposed method in size, shape and direction. Quantitative evaluation results reveal the excellent accuracy of the blur kernel model.

• Journal of Biomedical Engineering Research
• /
• v.40 no.2
• /
• pp.62-67
• /
• 2019

The purpose of this study was to evaluate the performance of deep neural network model in order to determine whether there is a risk factor for coronary artery disease based on the cardiac variation parameter. The study used unidentifiable 297 data to evaluate the performance of the model. Input data consists of heart rate parameters, which are SDNN (standard deviation of the N-N intervals), PSI (physical stress index), TP (total power), VLF (very low frequency), LF (low frequency), HF (high frequency), RMSSD (root mean square of successive difference) APEN (approximate entropy) and SRD (successive R-R interval difference), the age group and sex. Output data are divided into normal and patient groups, and the patient group consists of those diagnosed with diabetes, high blood pressure, and hyperlipidemia among the various risk factors that can cause coronary artery disease. Based on this, a binary classification model was applied using Deep Neural Network of deep learning techniques to classify normal and patient groups efficiently. To evaluate the effectiveness of the model used in this study, Kernel SVM (support vector machine), one of the classification models in machine learning, was compared and evaluated using same data. The results showed that the accuracy of the proposed deep neural network was train set 91.79% and test set 85.56% and the specificity was 87.04% and the sensitivity was 83.33% from the point of diagnosis. These results suggest that deep learning is more efficient when classifying these medical data because the train set accuracy in the deep neural network was 7.73% higher than the comparative model Kernel SVM.

• Safety and Health at Work
• /
• v.14 no.1
• /
• pp.93-99
• /
• 2023

Background: Effort-reward imbalance (ERI) and overcommitment at work have been associated poorer mental health. However, nonlinear and nonadditive effects have not been investigated previously. Methods: The association between effort, reward, and overcommitment with odds of poorer mental health was examined among a sample of 68 formal United States waste workers (87% male). Traditional, logistic regression and Bayesian Kernel machine regression (BKMR) modeling was conducted. Models controlled for age, education level, race, gender, union status, and physical health status. Results: The traditional, logistic regression found only overcommitment was significantly associated with poorer mental health (IQR increase: OR = 6.7; 95% CI: 1.7 to 25.5) when controlling for effort and reward (or ERI alone). Results from the BKMR showed that a simultaneous IQR increase in higher effort, lower reward, and higher overcommitment was associated with 6.6 (95% CI: 1.7 to 33.4) times significantly higher odds of poorer mental health. An IQR increase in overcommitment was associated with 5.6 (95% CI: 1.6 to 24.9) times significantly higher odds of poorer mental health when controlling for effort and reward. Higher effort and lower reward at work may not always be associated with poorer mental health but rather they may have an inverse, U-shaped relationship with mental health. No interaction between effort, reward, or overcommitment was observed. Conclusion: When taking into the consideration the relationship between effort, reward, and overcommitment, overcommitment may be most indicative of poorer mental health. Organizations should assess their workers' perceptions of overcommitment to target potential areas of improvement to enhance mental health outcomes.

• Land and Housing Review
• /
• v.15 no.1
• /
• pp.57-75
• /
• 2024

To create dynamic and bustling urban environments, a diverse array of commercial facilities is indispensable. These facilities are recognised as pivotal in attracting and accommodating a larger floating population, thereby suggesting that a greater diversity of commercial establishments fosters heightened consumer expenditure. With this premise, our study endeavours to explore the influence of commercial facility diversity on the Consumer Centre Index. Focused on the temporal context of 2021 and the spatial context of Seoul, our analysis utilizes the Consumer Centre Index, derived from Kernel Density analysis, as the dependent variable. Independent variables encompass factors reflecting commercial attributes and urban characteristics. Employing spatial regression analysis at the administrative district level, we discern that the clustering of similar industries exerts a more pronounced positive effect on consumer activation compared to the clustering of disparate industries. Additionally, the findings underscore the importance of concentrating industries that bolster consumer activation. Anticipated outcomes of this study include insights beneficial for optimizing commercial facility location policies within the consumer market.

• Journal of Intelligence and Information Systems
• /
• v.17 no.4
• /
• pp.241-254
• /
• 2011

Financial time-series forecasting is one of the most important issues because it is essential for the risk management of financial institutions. Therefore, researchers have tried to forecast financial time-series using various data mining techniques such as regression, artificial neural networks, decision trees, k-nearest neighbor etc. Recently, support vector machines (SVMs) are popularly applied to this research area because they have advantages that they don't require huge training data and have low possibility of overfitting. However, a user must determine several design factors by heuristics in order to use SVM. For example, the selection of appropriate kernel function and its parameters and proper feature subset selection are major design factors of SVM. Other than these factors, the proper selection of instance subset may also improve the forecasting performance of SVM by eliminating irrelevant and distorting training instances. Nonetheless, there have been few studies that have applied instance selection to SVM, especially in the domain of stock market prediction. Instance selection tries to choose proper instance subsets from original training data. It may be considered as a method of knowledge refinement and it maintains the instance-base. This study proposes the novel instance selection algorithm for SVMs. The proposed technique in this study uses genetic algorithm (GA) to optimize instance selection process with parameter optimization simultaneously. We call the model as ISVM (SVM with Instance selection) in this study. Experiments on stock market data are implemented using ISVM. In this study, the GA searches for optimal or near-optimal values of kernel parameters and relevant instances for SVMs. This study needs two sets of parameters in chromosomes in GA setting : The codes for kernel parameters and for instance selection. For the controlling parameters of the GA search, the population size is set at 50 organisms and the value of the crossover rate is set at 0.7 while the mutation rate is 0.1. As the stopping condition, 50 generations are permitted. The application data used in this study consists of technical indicators and the direction of change in the daily Korea stock price index (KOSPI). The total number of samples is 2218 trading days. We separate the whole data into three subsets as training, test, hold-out data set. The number of data in each subset is 1056, 581, 581 respectively. This study compares ISVM to several comparative models including logistic regression (logit), backpropagation neural networks (ANN), nearest neighbor (1-NN), conventional SVM (SVM) and SVM with the optimized parameters (PSVM). In especial, PSVM uses optimized kernel parameters by the genetic algorithm. The experimental results show that ISVM outperforms 1-NN by 15.32%, ANN by 6.89%, Logit and SVM by 5.34%, and PSVM by 4.82% for the holdout data. For ISVM, only 556 data from 1056 original training data are used to produce the result. In addition, the two-sample test for proportions is used to examine whether ISVM significantly outperforms other comparative models. The results indicate that ISVM outperforms ANN and 1-NN at the 1% statistical significance level. In addition, ISVM performs better than Logit, SVM and PSVM at the 5% statistical significance level.

• Atmosphere
• /
• v.25 no.2
• /
• pp.375-387
• /
• 2015

A Selected Multi-model CONsensus (SMCON) technique was developed and verified for the tropical cyclone track forecast in the western North Pacific. The SMCON forecasts were produced by averaging numerical model forecasts showing low 70% latest 6 h prediction errors among 21 models. In the homogeneous comparison for 54 tropical cyclones in 2013 and 2014, the SMCON improvement rate was higher than the other forecasts such as the Non-Selected Multi-model CONsensus (NSMCON) and other numerical models (i.e., GDAPS, GEPS, GFS, HWRF, ECMWF, ECMWF_H, ECMWF_EPS, JGSM, TEPS). However, the SMCON showed lower or similar improvement rate than a few forecasts including ECMWF_EPS forecasts at 96 h in 2013 and at 72 h in 2014 and the TEPS forecast at 120 h in 2013. Mean track errors of the SMCON for two year were smaller than the NSMCON and these differences were 0.4, 1.2, 5.9, 12.9, 8.2 km at 24-, 48-, 72-, 96-, 120-h respectively. The SMCON error distributions showed smaller central tendency than the NSMCON's except 72-, 96-h forecasts in 2013. Similarly, the density for smaller track errors of the SMCON was higher than the NSMCON's except at 72-, 96-h forecast in 2013 in the kernel density estimation analysis. In addition, the NSMCON has lager range of errors above the third quantile and larger standard deviation than the SMCON's at 72-, 96-h forecasts in 2013. Also, the SMCON showed smaller bias than ECMWF_H for the cross track bias. Thus, we concluded that the SMCON could provide more reliable information on the tropical cyclone track forecast by reflecting the real-time performance of the numerical models.