• KIPS Transactions on Software and Data Engineering
• /
• v.7 no.10
• /
• pp.387-396
• /
• 2018

Stock price prediction has been a difficult problem to solve. There have been many studies to predict stock price scientifically, but it is still impossible to predict the exact price. Recently, a variety of types of cryptocurrency has been developed, beginning with Bitcoin, which is technically implemented as the concept of distributed ledger. Various approaches have been attempted to predict the price of cryptocurrency. Especially, it is various from attempts to stock prediction techniques in traditional stock market, to attempts to apply deep learning and reinforcement learning. Since the market for cryptocurrency has many new features that are not present in the existing traditional stock market, there is a growing demand for new analytical techniques suitable for the cryptocurrency market. In this study, we first collect and process seven cryptocurrency price data through Bithumb's API. Then, we use the gradient boosting model, which is a data-driven learning based machine learning model, and let the model learn the price data change of cryptocurrency. We also find the most optimal model parameters in the verification step, and finally evaluate the prediction performance of the cryptocurrency price trends.

• Korean Journal of Plant Taxonomy
• /
• v.42 no.4
• /
• pp.247-252
• /
• 2012

Seed morphology in the subtribe Agrimoniinae (Rosaceae) was examined using scanning electron microscopy to identify distinct characters and evaluate their evolution in a phylogenetic framework for five genera in the subtribe: Agrimonia L., Aremonia Neck. ex Nestl., Hagenia J.F. Gmel., Leucosidea Eckl. & Zeyh., and Spenceria Trimen. All genera have one or two mature achenes in a fruiting hypanthium. In the seed coats, the cell shape, size, wall features, and sculpturing vary across genera. Of most significance is the presence of papillae structures in both Agrimonia and Aremonia. Through the mapping of papillae features onto phylogenetic trees, either one or two changes in seed coats are hypothesized. The phylogenetic tree inferred from four nuclear and six chloroplast regions of sequence data suggests that at least two steps of papillae sculpturing on seed coats are required. On the other hand, in the phylogenetic tree of a low-copy nuclear gene, one independent evolutionary step is postulated to explain the current character states. In the latter hypothesis, the seed coat sculpturing also supports a monophyletic relationship for cosmopolitan Agrimonia and European endemic Aremonia. The seed coat sculpturing provides valuable information for inferring phylogenetic relationships at the generic level in the subtribe Agrimoniinae.

• Journal of the Society of Naval Architects of Korea
• /
• v.54 no.6
• /
• pp.476-484
• /
• 2017

The purpose of this paper is to verify the structural integrity of a region with numerous penetration-holes in offshore structures such as semi-submersible rig and FPSO. In order to effectively check the yielding and buckling strength of plate members with penetration-holes, a screening analysis program was developed with the FE analysis tool to generate fine meshed model using the theoretical and analysis methods. When a hole is appeared in the plate structure members, the flow of stress is altered such that concentrations of stress form near the hole. Stress concentrations are of concern during both preliminary and detail design and need to be addressed from the perspectives of strength. To configure the geometrical shape, very fine meshed FE analysis is needed as the most accurate method. However, this method is practically impossible to apply for the strength verifications for all perforated plates. In this paper, screening analysis method was introduced to reduce analysis tasks prior to detailed FE analysis. This method is applied to not only the peak stress calculation combined stress concentration factor with nominal stress but also nominal equivalent stress calculation considering cutout effects. The areas investigated by very fine meshed analysis were to be chosen through screening analysis without any reinforcements for penetration-holes. If screening analysis results did not satisfy the acceptance criteria, direct FE analysis method as the 2nd step approach were applied with one of the coarse meshed model considering hole or with the very fine meshed model considering the hole shape and size. In order to effectively perform the local fine meshed analysis, automatic model generating program was developed based on the MSC/PATRAN which is pre-post FE analysis program. Buckling strength was also evaluated by Common Structure Rule (CSR) adopted by IACS as the stress obtained from very fine meshed FE analysis. Due to development of the screening analysis program and automatic FE modeling program, it was able to reduce the design periods and structural analysis costs.

• Proceedings of the Materials Research Society of Korea Conference
• /
• 2011.05a
• /
• pp.8.1-8.1
• /
• 2011

Nanocrystalline titanium dioxide ($TiO_2$) materials have been widely used as an electron collector in DSSC. This is required to have an extremely high porosity and surface area such that the dye can be sufficiently adsorbed and be electronically interconnected, resulting in the generation of a high photocurrent within cells. In particular, their geometrical structures and crystalline phase have been extensively investigated as important issues in improving its photovoltaic efficiency. In this study, we present a new strategy to fabricate a photoelectrode having a periodic structured $TiO_2$ film templated from 1D or 3D polystyrene (PS) microspheres array. Monodisperse PS spheres of various radiuses were used for colloidal array on FTO glasses and two types of photoelectrode structures with different $TiO_2$ materials were investigated respectively. One is the igloo-shaped electrode prepared by $TiO_2$ deposition by RF-sputtering onto 2D microsphere-templated substrates. At the interface between the film and substrate, there are voids formed by the decomposition of PS microspheres during the calcination step. These holes might be expected to play the predominant roles as scattering spherical voids to promote a light harvesting effect, a spacious structure for electrolytes with higher viscosity and effective paths for electron transfer. Additionally the nanocrystalline $TiO_2$ phase prepared by the RF-sputtering method was previously reported to improve the electron drift mobility within $TiO_2$ electrodes. This yields solar cells with a cell efficiency of 2.45% or more at AM 1.5 illumination, which is a very remarkable result, considering its $TiO_2$ electrode thickness (<2 ${\mu}m$). This study can be expanded to obtain higher cell efficiency by higher dye loading through the increase of surface area or multi-layered stacking. The other is the inverse opal photonic crystal electrode prepared by titania particles infusion within 3D colloidal arrays. To obtain the enlargement of ordered area and high quality of crystallinity, the synthesis of titania particles coated with a organic thin layer were applied instead of sol-gel process using the $TiO_2$ precursors. They were dispersed so well in most solvents without aggregates and infused successfully within colloidal array structures. This ordered mesoporous structure provides the large surface area leading to the enough adsorption of dye molecules and have an light harvesting effect due to the photonic band gap properties (back-and-forth reflection effects within structures). A major advantage of this colloidal array template method is that the pore size and its distribution within $TiO_2$ photoelectrodes are determined by those of latex beads, which can be controlled easily. These materials may have promising potentials for future applications of membrane, sensor and so on as well as solar cells.

• Korean Journal of Materials Research
• /
• v.21 no.1
• /
• pp.8-14
• /
• 2011

To improve the chemical stability of metal, the ceramic coatings on metallic materials have attracted interest from many researchers due to the chemical inertness of ceramic materials. To endure strong acids, SiOC coating on metal substrate was carried out by dip coating method using 20wt% polyphenylcarbosilane solution; SiC powder was added to the solution at 10wt% and 15wt% to improve the mechanical properties and to prevent cracks of the film. Thermal oxidation as a curing step was carried out at $200^{\circ}C$ for crosslinking of the polyphenylcarbosilane, and the coating samples were pyrolysized at $800^{\circ}C$ under argon to convert the polyphenylcarbosilane to SiOC film. The thicknesses of the SiOC coating films were $2.36{\mu}m$ and $3.16{\mu}m$. The quantities of each element were measured as $SiO_{1.07}C_{6.33}$ by EPMA, and it can be confirmed that the SiOC film from polyphenylcarbosilane was formed in a manner that was carbon rich. The hardness of the SiOC film was found to be 3.2Gpa through nanoindentor measurement. No defect including cracks appeared in the SiOC film. The weight loss of the SiOC coated stainless steel was within 2% after soaking in 10% HCl solution at $80^{\circ}C$ for one week. From these results, SiOC coating shows good potential for application to protect against severe chemical corrosion of stainless steel.

• Journal of Korean Society of Transportation
• /
• v.26 no.5
• /
• pp.163-174
• /
• 2008

The conventional four-step travel demand model is still widely used as the state-of-practice in most transportation planning agencies even though it does not provide reliable estimates of travel demand. In order to improve the accuracy of travel demand estimation, implementing an alternative approach would be critical as much as acquiring reliable socioeconomic and travel data. Recently, the role of travel demand model is diverse to satisfy the needs of microscopic analysis regarding various policies of travel demand management and traffic operations. In this context, the activity-based approach for travel demand estimation is introduced and a case study of developing a spatial-temporal activity presence-based approach that estimates travel demand through forecasting number of people present at certain place and time is accomplished. Results show that the spatial-temporal activity presence-based approach provides reliable estimates of both number of people present and trips actually people made. It is expected that the proposed approach will provide better estimates and be used in not only long-term transport plans but short-term transport impact studies with respect to various transport policies. Finally, in order to introduce the spatial-temporal activity presence-based approach, the data such as activity-based travel diary and land use based on geographic information system are essential.

• Journal of the Korean Society of Hazard Mitigation
• /
• v.10 no.5
• /
• pp.149-158
• /
• 2010

Recent rainfall patterns in Korea show that both of the total amount of rainfall and the total number of heavy rain days have been increased. Therefore, the damage resulted from flood disaster has been dramatically increased in Korea. The purpose of the present study is to analyze flooding in an urban area using SWMM linked with FLUMEN. The study area is Suyeong-Mangmi lowland area, Busan, Korea. Suyeong-Mangmi lowland area have been a flooding hazard zone since 1995. The last flooding cases of this area occurred on July 7th and 16th, 2009, and the later flooding case was analyzed in this study. The first step of computation is calculating flow through storm sewers using the urban runoff simulation model of SWMM. The flooding hydrographs are used in the inundation analysis model of FLUMEN. The results of inundation analysis were compared with the real flooding situation of the study area. The real maximum inundation depth was guessed by 1.0 m or more on July 16th. The computation yields the maximum inundation depth of 1.2 m and the result was somewhat overestimated. The errors may be resulted from the runoff simulation and incapability of simulation using FLUMEN for flow into buildings. The models and procedures used in this study can be applied to analysis of flooding resulted from severe rainfall and insufficiency of drainage capacity.

• Journal of the Korean Chemical Society
• /
• v.4 no.1
• /
• pp.18-26
• /
• 1957

There are two problems to be solved by our efforts in the enamel frit. One is how we can cover the enamel frit thin with complete milk white as possible, and the other is how it can be, made resistant for chemicals than before one. The frit which can solved the two problems just mentioned above is titanium enamel frit. This frit has been developed in America after War Ⅱ, and now the research for concerning antimony frit into titanium frit is under development entirely. In order to develope the enamel industry in Korea, it is urgent problem to convert antimony frit into titanium frit. By the way the titanium frit is emulsified titanium oxide crystal which made through reheating the supersaturated solution of titanium oxide in the basis of glass. Unfortunately, there are many obscure points in active fact or which influence on its composition and characteristics yet. However, this task was tried for the first in Korea. As first step, the test was carried on the reference books, and we can be possible convert antimony frit into titanium frit as a result of this experiment. As a conclusion, for the purpose of developing the enamel industry in Korea, we studied that the research for converting antimony enamel frit which has been used popularly into titanium enamel frit which is more economic and resistant for chemicals. As a result of experiments, the following points concerning with titanium frit have become clearly. 1. It is better when the composition of titanium enamel frit has as following table.Man Duck San Silica 24 An Yang Feldspar 20 Borax 28 Sodium Nitrate 4 Cryolite 7 Calcium Carbonate 3.6∼1 Titanium Oxide 10 Calcium phosphate 0 ∼3.2 Calcium Fluoride 0∼1.8 Antimony Oxide 0∼0.5 2. The amount of $TiO_2$, to be added is $10%\;to\;12{%,\;CaF_2\;is\;under\;1.8%,\;P_2O_5\;is\;under\;1.6%,\;Sb_2O_3\;is\;under\;0.5%$. 3. In the titanium frit, the limit of iron oxide amount to be included is under 0. 5%. 4. Comparing the titanium enamel frit with antimony enamel frit not only the titanium frit can be savely 20.6% in the price of raw materials, but one time of glazing and heating process is omitted in each case, and it is known the titanium frit is more resistant for chemicals than antimony frit.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.9 no.3
• /
• pp.146-155
• /
• 2006

A handful of previous studies have attempted to integrate socioeconomic data and remotely sensed data for urban quality of life assessment with their spatial dimension in a zonal unit. However, such a zone-based approach not only has the unrealistic assumption that all attributes of a zone are uniformly spatially distributed throughout the zone, but also has resulted in serious methodological difficulties such as the modifiable areal unit problem and the incompatibility problem with environmental data. An alternative to the zone-based approach can be a pixel-based approach which gets its spatial dimension through a pixel. This paper proposes a pixel-based approach to linking remotely sensed data with socioeconomic data in GIS for urban quality of life assessment. The pixel-based approach uses dasymetric mapping and spatial interpolation to spatially disaggregate socioeconomic data and integrates remotely sensed data with spatially disaggregated socioeconomic data for the quality of life assessment. This approach was implemented and compared with a zone-based approach using a case study of Fulton County, Georgia. Results indicate that the pixel-based approach allows for the calculation of a microscale indicator in the urban quality of life assessment and facilitates efficient data integration and visualization in the assessment although it costs an intermediate step with more processing time such as the disaggregation of zonal data. The results also demonstrate that the pixel-based approach opens up the potential for the development of new database and increased analytical capabilities in urban analysis.

• Journal of Digital Convergence
• /
• v.12 no.6
• /
• pp.203-213
• /
• 2014

The purpose of this research is to deduce the effects factor of R&D output and outcome in the aspect of the researcher's intellectual property capacity, and conduct an empirical analysis to see if there is intermediated effect between cooperation of a patent attorney and output. I applied three step, input (researcher's intellectual property capacity and intellectual property environment), output (number of patent and SCI paper), outcome (number of technology transfer, profit), and analyzed the effects of researcher's intellectual property factors on output and outcome through structural equation model. The results are as follows. First, there's no cooperation between a researcher and patent attorney, the researcher's knowledge on patent specification had direct effect on the output. Second, a researcher and an patent attorney cooperate, the cooperative activities have complete mediation effect on the output, and the researcher's will to commercialize the technology has direct effects on the output. Third, the intellectual property environment of the research organization gives direct effects on the researcher's intellectual property capacity, and gives indirect effects on the outcome. this study suggest that it is important to cooperate with external experts in order to excellent R&D outcome.