• Journal of Korean Society for Atmospheric Environment
• /
• v.32 no.5
• /
• pp.501-512
• /
• 2016

The study evaluated methods to measure condensable fine particles in flue gases and measured particulate matter by fuel and material to get precise concentrations and quantities. As a result of the method evaluation, it is required to improve test methods for measuring Condensable Particulate Matter (CPM) emitted after the conventional Filterable Particulate Matter (FPM) measurement process. Relative Standard Deviation (RSD) based on the evaluated analysis process showed that RSD percentages of FPM and CPM were around 27.0~139.5%. As errors in the process of CPM measurement and analysis can be caused while separating and dehydrating organic and inorganic materials from condensed liquid samples, transporting samples, and titrating ammonium hydroxide in the sample, it is required to comply with the exact test procedures. As for characteristics of FPM and CPM concentrations, CPM had about 1.6~63 times higher concentrations than FPM, and CPM caused huge increase in PM mass concentrations. Also, emission concentrations and quantities varied according to the characteristics of each fuel, the size of emitting facilities, operational conditions of emitters, etc. PM in the flue gases mostly consisted of CPM (61~99%), and the result of organic/inorganic component analysis revealed that organic dusts accounted for 30~88%. High-efficiency prevention facilities also had high concentrations of CPM due to large amounts of $NO_x$, and the more fuels, the more inorganic dusts. As a result of comparison between emission coefficients by fuel and the EPA AP-42, FPM had lower result values compared to that in the US materials, and CPM had higher values than FPM. For the emission coefficients of the total PM (FPM+CPM) by industry, that of thermal power stations (bituminous coal) was 71.64 g/ton, and cement manufacturing facility (blended fuels) 18.90 g/ton. In order to estimate emission quantities and coefficients proper to the circumstances of air pollutant-emitting facilities in Korea, measurement data need to be calculated in stages by facility condition according to the CPM measurement method in the study. About 80% of PM in flue gases are CPM, and a half of which are organic dusts that are mostly unknown yet. For effective management and control of PM in flue gases, it is necessary to identify the current conditions through quantitative and qualitative analysis of harmful organic substances, and have more interest in and conduct studies on unknown materials' measurements and behaviors.

• 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.

• Journal of The Korean Association For Science Education
• /
• v.35 no.3
• /
• pp.403-417
• /
• 2015

The purpose of this study was to investigate the effectiveness of integrated education research conducted in Korea and to propose a meaningful discussion for further research. Among the studies conducted for last three years, the relevant 161 research articles were selected, and 236 effect sizes were calculated. Effect sizes were analyzed with different dependant variables including creativity, problem solving ability, academic achievement, inquiry skills, creative personality, scientific attitude, and interests. In addition, effect sizes with different moderating variables, such as characteristics of subjects, sample sizes, class types, core disciplines and publication types, were compared. The results are as follows: The overall effect size of integrated education program produced a huge effect (effect size=0.88, U3=81.06%). Integrated education program showed the highest effect size on scientific attitude among other dependant variables. However, all of the other dependant variables represented more than medium size effect size. Integrated program proved to be more effective on kindergarten pupils and gifted students compared to other school levels and regular students. The effect size for group of less then thirty students were larger than other groups. Programs implemented in after school hours were more effective than in regular school hours. Considering the core subject of program, arts-centered integrated programs showed the largest effect size, while all the others showed above medium effect sizes. Finally, doctoral dissertation showed the highest effect size compared to master's thesis and academic journal articles. Conclusions and recommendations for further research were provided.

• Economic and Environmental Geology
• /
• v.40 no.5
• /
• pp.675-689
• /
• 2007

Kilauea volcano's summit area was formed by continuous ind/or sporadic eruption activities for several hundreds years. In this study, we mainly focused on the trace elements characteristics through systematic sample rocks erupted from 1790 to September of 1982. Under the microscope it can be observed some main minerals such as olivine, clinopyroxene. and plagioclase with minor opaque minerals including Cr-spinel and ilmenite. Zr, V, Y, Ti elements show incompatible activities with MgO while Ni, Cr, Co elements show highly compatible properties. Elements like as Ba, Rb, Th, Sr, Nd are highly incompatible to show positive trends with $K_2O$. In the REE diagram LREE is more enriched than HREE suggesting typical Oceanic Island Basalt(OIB) type. It can be suggested that Sr have an effect on the fractionation of plagioclase from the kink in the $K_2O$ variation diagram. Y/Ho ratio diagram shows there was no fluids effect in the historical Kilauea volcano but Zr/Hf ratio diagram shows a significant difference between Kilauea lavas and PuuOo lavas. There are distinctive changes of trace element contents showing in particular abrupt changes of temporal variations between 1924 and 1954. Moreover, PuuOo lavas which had been erupted since 1983 follow these decreasing trends of trace element variation. Therefore, it is strongly suggested that these abrupt changes of trace elements trends result from the huge collapse geological event which formed Halemaumau crater in 1924 causing contamination effects of crustal contents into magma chamber and from the changes of parental magma composition injected into Kilauea volcano's summit magma reservoir.