• Proceedings of the Korea Information Processing Society Conference
• /
• 2010.04a
• /
• pp.884-887
• /
• 2010

최근 검색엔진의 효율성을 향상시키고 검색결과에 있어서 사용자들의 요구사항을 충족시키기 위한 연구들이 활발히 수행되고 있으며, 많은 방법론들이 제시되고 있다. 이는 방대한 정보 속에서 사용자의 검색 의도에 맞는 정보를 효과적으로 제공하는 것을 그 목표로 한다. 특히 본 논문에서는 검색하고자 하는 토픽별 사용자의 인적 속성들이 미치는 영향력을 기반으로 사용자간 소셜 관계 지수(SRV : Social Relationship Value)를 산정하는 방법을 제안한다. 소셜 관계 지수란 인간의 내재적인 특성을 수치로 산정한 것으로, 웹 사용자들에게 있어서는 검색 성향의 유사정도와 직결된다. 따라서 검색하고자 하는 토픽별 개인 성향의 유사정도를 수치로 부여하고 유사성이 높은 사람들의 검색 정보를 이용하면 사용자에 보다 만족된 검색결과를 제공할 수 있다. 본 연구에서는 구글 디렉터리(Google directory)의 정제된 각 토픽별 하위 범주(category)에 대해 선택 결과가 같은 사람들을 대상으로 인적 속성을 분석하고, 그 영향력을 가중치로 적용해 산정된 소셜 관계 지수와 사용자들의 검색 패턴을 비교 하였다. 그 결과 특정인을 기준으로 소셜 관계 지수가 높은 사람들의 검색 패턴이 매우 유사함을 확인 하였다. 이를 통해 토픽별 개인 간 연결 강도가 강할수록, 즉 유사성이 높은 사용자간에는 검색 패턴 또한 유사함을 검증 할 수 있었다.

• Journal of the Society of Disaster Information
• /
• v.19 no.4
• /
• pp.924-935
• /
• 2023

Purpose: The central aim of this study is to leverage machine learning techniques for the classification of Intrusion Detection System (IDS) data, with a specific focus on identifying the variables responsible for enhancing overall performance. Method: First, we classified 'R2L(Remote to Local)' and 'U2R (User to Root)' attacks in the NSL-KDD dataset, which are difficult to detect due to class imbalance, using seven machine learning models, including Logistic Regression (LR) and K-Nearest Neighbor (KNN). Next, we use the SHapley Additive exPlanation (SHAP) for two classification models that showed high performance, Random Forest (RF) and Light Gradient-Boosting Machine (LGBM), to check the importance of variables that affect classification for each model. Result: In the case of RF, the 'service' variable and in the case of LGBM, the 'dst_host_srv_count' variable were confirmed to be the most important variables. These pivotal variables serve as key factors capable of enhancing performance in the context of classification for each respective model. Conclusion: In conclusion, this paper successfully identifies the optimal models, RF and LGBM, for classifying 'R2L' and 'U2R' attacks, while elucidating the crucial variables associated with each selected model.

• Convergence Security Journal
• /
• v.18 no.4
• /
• pp.62-71
• /
• 2018

The rapid growth of virtual reality technology in the era of the 4th Industrial Revolution has accelerated scientification of combat training systems in addition to ICT(information and communications technology) in military field. Recently, research and development of simulators based on virtual reality have been actively conducted in order to solve sensitive issues such as increase of civil complaints due to the noise of a shooting range, prevention of shooting accident, and reduction of training cost. In this paper, we propose two key solutions: spatial synchronization method and modified point mass trajectory model with small angle approximation to overcome technical limitations of a current training simulator. A trainee who wears a haptic vest in a mixed reality environment built in MARS(medium-range assault rifle shooting simulator) is able to conduct not only precision shooting but also two-way engagement with virtual opponents. It is possible for trainee to receive more reliable evaluations in the MARS than an existing rifle simulator based on laser.

• Journal of Animal Science and Technology
• /
• v.47 no.4
• /
• pp.691-696
• /
• 2005

Suckling piglets must avoid positions with high air velocity because they don’t have hypodermic fat. Therefore this study examined the effects of air ejection on the behaviors of sows and their piglets. Compressed air was released for 5s at 5s intervals between floor and udder of sows only when they were standing or sitting. Sixteen multiparous crossbred(Landrace×Yorkshire) sows and their piglets were used. Behaviors of sixteen sows and their piglets were recorded for 4 days postpartum, using the LED lamp(wavelength:950nm), CCD camera(Samsung SDC-411), multiplexer(Samsung SDM-080), and time lapsed VCR(Samsung SRV-30). The videotapes were scanned every 30s to obtain an instantaneous behavioral sample. The sow’s standing and sitting rates between control group(CG) and air ejection group(AEG) were not significantly different(P＞0.05). This means that air ejection does not affect the behavior of sows. Frequency of the suckling piglets’ behaviors closely related to the crushing by sows was lower in AEG than in CG(P＜0.05). These results suggest that air ejection may be available for reduction of the crushing of suckling piglets by sows.

• Korean Journal of Veterinary Research
• /
• v.45 no.2
• /
• pp.245-250
• /
• 2005

This study was carried out to find out how space allowance affect the social behavior of Korean native cattle (Bos taurus coreanae) steers. Twelve Korean native cattle (Bos taurus coreanae) steers were used as subjects, each of which was 30-month-old and observation period was from June to July 2003. Five (T1) and seven (T2) steers were allotted to two pens of $5m{\times}10m$ in a building with slate roof and open sides respectively. They were fed at 09:00 h and 16:00 h, twice a day. The behaviors of steers were recorded from 06:00 h to 17:00 h, using two color CCD cameras (Samsung SDC-411, Korea), one B/W CCD cameras (Samsung SBC-340, Korea), one multiplexer (Samsung SDM-081, Korea) and a time lapse VCR (Samsung SRV-30, Korea). The behaviors of each steer were recorded every 2 min using an instantaneous point sampling method. While the mean percentage of time budget in WA of T1 was lower than that of T2 (p<0.05), the mean percentage of time budget in SF of T1 was higher than that of T2 (p<0.05). When it gets hot, steers in T1 rested from 10:00 h to 14:00 h when it gets cool, showing 40~80% of LD rate while steers in T2 rested from 12:00 h, when it very hot to 17:00 h, showing 20~50% of LD rate, which is relatively low. Steers in T1 were fed from 06:00 h to 08:00 h when it was cool and from 16:00 h to 18:00 h, showing 20~45% of EA rate while steers in T2 were fed from 08:00 h to 14:00 h when it was hot, showing 25~50% of EA rate. In conclusion, it turned out that the number of steers affected their social behavior, and T1 was better environment than T2 in terms of welfare.

• Journal of the Korean Institute of Gas
• /
• v.20 no.1
• /
• pp.13-22
• /
• 2016

This paper presents development of an appropriate procedure and flow chart to analyze shale gas production data obtained from a multi-fractured horizontal well according to flow characteristics in order to calculate an estimated ultimate recovery. Also, the technical considerations were proposed when a rate transient analysis was performed with field production data occurred to only $1^{st}$ transient flow. If production data show the $1^{st}$ transient flow from log-log and square root time plot analysis, production forecasting must be performed by applying different method as before and after of the end of $1^{st}$ linear flow. It is estimated by an area of stimulated reservoir volume which can be calculated from analysis results of micro-seismic data. If there are no bottomhole pressure data or micro-seismic data, an empirical decline curve method can be used to forecast production performance. If production period is relatively short, an accuracy of production data analysis could be improved by analyzing except the early production data, if it is necessary, after evaluating appropriation with near well data. Also, because over- or under-estimation for stimulated reservoir volume could take place according to analysis method or analyzer's own mind, it is necessary to recalculate it with fracture modeling, reservoir simulation and rate transient analysis, if it is necessary, after adequacy evaluation for fracture stage, injection volume of fracture fluid and productivity of producers.

• Tunnel and Underground Space
• /
• v.31 no.1
• /
• pp.41-51
• /
• 2021

The hydraulic fracturing developed to improve permeability of tight reservoir is one of key stimulation technologies for developing unconventional resources such as shale gas and deep geothermal energy. The experimental study was conducted to improve disadvantage of hydraulic fracturing which has simple fracture pattern and poor fracturing efficiency. The fracturing experiments was conducted for tight rock using various fracturing fluids, water, N2, and CO2 and the created fracture pattern and fracturing efficiency was analyzed depending on fracturing fluids. The borehole pressure increased rapidly and then made fractures for hydraulic fracturing with constant injection rate, however, gas fracturing shows slowly increased pressure and less fracture pressure. The 3D tomography technic was used to generate images of induced fracture using hydraulic and gas fracturing. The stimulated reservoir volume (SRV) was estimated increment of 5.71% (water), 12.72% (N2), and 43.82% (CO2) respectively compared to initial pore volume. In addition, permeability measurement was carried out before and after fracturing experiments and the enhanced permeability by gas fracturing showed higher than hydraulic fracturing. The fracture conductivity was measured by increasing confining stress to consider newly creating fracture and closing induced fracture right after fracturing. When the confining stress was increased from 2MPa to 10MPa, the initial permeability was decreased by 89% (N2) and 50% (CO2) respectively. This study shows that the gas fracturing makes more permeability enhancement and less reduction of induced fracture conductivity than hydraulic fracturing.