• Proceedings of the Korea Information Processing Society Conference
• /
• 2008.11a
• /
• pp.714-717
• /
• 2008

협업여과 추천기법에는 사용자 기반 협업여과와 아이템 기반 협업여과가 있으며, 절차는 유사도 측정, 이웃 선정, 예측값 생성 단계로 이루어진다. 유사도 측정 단계에는 유클리드 거리(Euclidean Distance), 코사인 유사도(Cosine Similarity), 피어슨 상관계수(Pearson Correlation Coefficient) 방법 등이 있고, 이웃 선정 단계에는 상관 한계치(Correlation-Threshold), 근접 N 이웃(Best-N-Neighbors) 방법 등이 있다. 마지막으로 예측값 생성 단계에는 단순평균(Simple Average), 가중합(Weighted Sum), 조정 가중합(Adjusted Weighted Sum) 등이 있다. 이처럼 협업여과 추천기법에는 다양한 기법들이 사용되고 있다. 따라서 본 논문에서는 사용자 기반 협업여과와 아이템 기반 협업여과 추천기법에 사용되는 유사도 측정 기법과 예측값 생성 기법의 최적화된 조합을 알아보기 위해 성능 실험 및 비교 분석을 하였다. 실험은 GroupLens의 MovieLens 데이터 셋을 활용하였고 MAE(Mean Absolute Error)값을 이용하여 추천기법을 비교 하였다. 실험을 통해 유사도 측정 기법과 예측값 생성 기법의 최적화된 조합을 찾을 수 있었고, 사용자 기반 협업여과와 아이템 기반 협업여과의 성능비교를 통해 아이템 기반 협업여과의 성능이 보다 우수했음을 확인 하였다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2001.10b
• /
• pp.82-84
• /
• 2001

정보여과 시스템은 사용자의 관심사를 정확하게 알아내야 하고(specialization), 시간에 따른 변화에 적응할 수 있어야 하며(adaptation), 사용자의 잠재적인 관심사를 발견하기 위해 새로운 도메인을 탐험할 수 있어야 한다(exploration). 본 논문에서는 온라인 뉴스 기사를 여과하여 사용자와 관련이 있는 뉴스 기사를 추천하는 뉴스 에이전트를 설계, 구현하고자 한다. Specialization, adaptation의 두 가지 요구사항을 충족시키기 위해 사용자의 관심사를 도메인별로 분리하고 각 도메인은 long-term과 short-term으로 나눈다. Exploration의 요구사항을 충족시키기 위해서는 카테고리 절차(crossover) 연산을 사용한다. 실험 결과, 사용자에 대한 사전 정보가 전혀 없는 상태에도 불구하고 빠른 적응능력을 보였다. long-term과 short-term의 분리는 사용자의 관심사에 급격한 변화가 일어난 후에도 시스템이 빠르게 적응할 수 있음을 보여주었다. 또한 카테고리 교차 연산을 통해 사용자의 새로운 관심사 탐험을 수행해 낼 수 있음을 보여주었다.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.16 no.1
• /
• pp.59-68
• /
• 2003

An optimization Program is developed to produce new topologies of plane structures under multiload case. A four-node finite element is used in the response analysis to reduce the computation time and to ultimately achieve practical topology optimization. The bilinear finite element is prone to produce chequer-boarding phenomenon and a simple filtering process is therefore adopted. An artificial material model is employed to represent the structural material and the resizing algorithm based on the optimality criteria is adopted to update the material density parameter during optimization process. With newly developed optimization program, the comparison study has been made between single and multiload cases and its results are described in this paper. From numerical results, it appears that multiload case should be considered to achieve the practical topology optimization.

• Proceeding of KASS Symposium
• /
• 2006.05a
• /
• pp.233-238
• /
• 2006

This paper provides a new topology optimization technique which is intended to maximize the fundamental frequency with simultaneous consideration of other natural frequencies in the form of multi-frequency problems. The modal strain energy is considered as the objective function to be minimized and the initial volume of structures is used as the constraint function. The resizing algorithm based on the optimality criteria is adopted to update the hole size existing inside the material. From numerical tests, the proposed technique is found to be very effective to maximize the fundamental frequency of the structure and it can also successfully consider several higher mode effects into the optimum topology of structure through the introduction of weights.

• Journal of Korean Association for Spatial Structures
• /
• v.5 no.3 s.17
• /
• pp.83-89
• /
• 2005

This paper provides the results of the investigation on the optimum topology of simply supported deep beam structures with multi-point load cases. In this study, the strain energy to be minimized is considered as the objective function and the initial volume of structures is used as the constraint function. The resizing algorithm based on the optimality criteria is adopted to update the hole size existing inside the material. In this study, the sensitivities of topology optimization parameters to the optimum topology of the deep bean structures is investigated and also the effect of filtering process on the optimum topology is thoroughly tested. From numerical tests, the optimum topology of the deep beam is closely related with the optimization parameters used in the iteration and the filtering process play important role in order to find the optimum topology of the deep beam.

• Journal of Korean Society of Environmental Engineers
• /
• v.37 no.12
• /
• pp.689-695
• /
• 2015

A biofilm filtration for the removal of gaseous pollutants has been recognized as a process with a complex interaction between the gas flow characteristics and the process operating variables. This study aims to develop an one dimensional dynamic numerical model which can be utilized as a tool for the analysis of biofilm filtration process operated in plug flow mode. Since, in a plug flow system, minor environmental changes in a gaseous unit process cause a drastic change in reaction and the interaction between the pollutants is an influencing factor, plug flow system was generalized in developing the model. For facilitation of the model development, dispersion was simplified based on the principles of material balance. Several reactions such as competition, escalation, and control between the pollutants were included in the model. The applicability of the developed model was evaluated by taking the calibration and verification steps on the experimental data performed for the removal of BTX at both low and high flow concentration. The model demonstrated a correlation coefficient ($R^2$) greater than 0.79 under all the experimental conditions except for the case of toluene at high flow condition, which suggested that this model could be used for the generalized gaseous biofilm plug flow filtration system. In addition, this model could be a useful tool in analyzing the design parameters and evaluating process efficiency of the experiments with substantial amount of complexity and diversity.

• Proceedings of the Korean Nuclear Society Conference
• /
• 1997.10a
• /
• pp.387-391
• /
• 1997

냉각재중의 방사능을 띤 성분 중에는 이온교환기에서 제거가 가능한 이온성분과 함께 필터에 의해서 주로 제거되는 입자성 물질로 존재한다. 운전중의 냉각재내 방사성 부식생성물의 물리적 조성 분포 측정 결과에 따르면 90%이상이 0.45$mu extrm{m}$필터에 의해 제거되는 입자성 물질로 구성되어 있다. 이로 인해 새수지 충전후 얼마 사용하지 않은 탈염기의 제염계수가 탈염기에서 완벽한 제거가 어려운 입자성 부식생성물로 인해 10이하를 나타낼 수 있다. 1차계통에 쓰이는 수지의 성능검사를 위해 사용하고 있는 현재의 제염계수 측정법은 다음과 같은 두가지 이유로 완벽하지 않음을 알 수 있다. 첫째, 냉각재중의 방사능을 띤 성분중에는 이온교환기에서 제거가 가능한 이온성분과 함께 필터에 의해 제거되는 입자성 물질도 함께 존재하므로 탈염기의 제염계수 측정 절차는 입자성 물질을 배제한 후 측정해야 하며, 특히 수치 교체를 결정하기 위한 제염계수 측정시에는 여과된 여액으로 방사능 농도를 측정하는 것이 바람직하다. 둘째 운전중인 냉각재의 시료중에는 핵분열 수율이 높고 핵연료봉 손상부위로 유출이 용이한 불활성 기체핵종들이 많이 존재하며, 탈염기 후단에서 채취한 시료중에도 많이 존재하고, 시료 이송과 방사능 측정동안의 짧은 시간동안에도 계속 붕괴반응함으로서 새로 생긴 핵종으로 인해 마치 탈염기의 제거능이 낮은 것으로 오판될 수 있다. 이러한 측정 오차인자를 고려하여야 1차계통 탈염기의 교환능력을 정확히 판정할 수 있다.

• Membrane Journal
• /
• v.16 no.1
• /
• pp.39-50
• /
• 2006

Porous chitosan and chitin membranes were prepared by using silica particles as porogen. Membrane preparation was achieved via the following three steps: (1) chitosan film formation by casting an chitosan solution containing silica particles, (2) preparation of porous chitosan membrane by dissolving the silica particles by immersing the film into an alkaline solution and (3) preparation of porous chitin membrane by acetylation of chitosan membrane with acetic anhydride. The optimum preparation conditions which could provide a chitosan and chitin membranes with good mechanical strength and adequate pure water flux were determined. To allow protein affinity, a reactive dye (Cibacron Blue 3GA) was immobilized on porous chitosan membrane. Binding capacities of affinity chitosan and chitin membranes for protein and enzyme were determined by the batch adsorption experiments of BSA protein and lysozyme enzyme. The maximum binding capacity of affinity chitosan membrane for BSA protein is about 22 mg/mL, and that of affinity chitin membrane for lysozyme enzyme is about 26 mg/mL. Those binding capacities are about $several{\sim}several$ tens times larger than those of chitosan and chitin-based hydrogel beads. Those results suggest that the porous chitosan and chitin membranes are suitable in affinity filtration chromatography for large scale separation of proteins.

• Journal of KIISE:Software and Applications
• /
• v.27 no.3
• /
• pp.290-301
• /
• 2000

Most morphological analysis methods require repetitive procedures of input character code conversion, segmentation and lemmatization of constituent morphemes, filtering of candidate results through looking up lexicons, which causes run-time inefficiency. To alleviate such problem of run-time inefficiency, many systems have introduced the notion of 'pre-analysis' of words. However, this method based on pre-analysis dictionary of surface also has a critical drawback in its practical application because the size of the dictionaries increases indefinite to cover all words. This paper hybridizes both extreme approaches methodologically to overcome the problems of the two, and presents a method of morphological analysis based on pre-analysis of partial words. Under such hybridized scheme, most computational overheads, such as segmentation and lemmatization of morphemes, are shifted to building-up processes of the pre-analysis dictionaries and the run-time dictionary look-ups are greatly reduced, so as to enhance the run-time performance of the system. Moreover, additional computing overheads such as input character code conversion can also be avoided because this method relies upon no graphemic processing.

• KIPS Transactions on Computer and Communication Systems
• /
• v.2 no.6
• /
• pp.279-290
• /
• 2013

In digital forensics log files have been stored as a form of large data for the purpose of tracing users' past behaviors. It is difficult for investigators to manually analysis the large log data without clues. In this paper, we propose a text mining technique for extracting intrusion logs from a large log set to recommend reliable evidences to investigators. In the training stage, the proposed method extracts intrusion association words from a training log set by using Apriori algorithm after preprocessing and the probability of intrusion for association words are computed by combining support and confidence. Robinson's method of computing confidences for filtering spam mails is applied to extracting intrusion logs in the proposed method. As the results, the association word knowledge base is constructed by including the weights of the probability of intrusion for association words to improve the accuracy. In the test stage, the probability of intrusion logs and the probability of normal logs in a test log set are computed by Fisher's inverse chi-square classification algorithm based on the association word knowledge base respectively and intrusion logs are extracted from combining the results. Then, the intrusion logs are recommended to investigators. The proposed method uses a training method of clearly analyzing the meaning of data from an unstructured large log data. As the results, it complements the problem of reduction in accuracy caused by data ambiguity. In addition, the proposed method recommends intrusion logs by using Fisher's inverse chi-square classification algorithm. So, it reduces the rate of false positive(FP) and decreases in laborious effort to extract evidences manually.