• Journal of Intelligence and Information Systems
• /
• v.16 no.1
• /
• pp.93-116
• /
• 2010

Ensemble approach is applied to the detection modeling of illegal cash accommodation (ICA) that is the well-known type of fraudulent usages of credit cards in far east nations and has not been addressed in the academic literatures. The performance of fraud detection model (FDM) suffers from the imbalanced data problem, which can be remedied to some extent using an ensemble of many classifiers. It is generally accepted that ensembles of classifiers produce better accuracy than a single classifier provided there is diversity in the ensemble. Furthermore, recent researches reveal that it may be better to ensemble some selected classifiers instead of all of the classifiers at hand. For the effective detection of ICA, we adopt ensemble size reduction technique that prunes the ensemble of all classifiers using accuracy and diversity measures. The diversity in ensemble manifests itself as disagreement or ambiguity among members. Data imbalance intrinsic to FDM affects our approach for ICA detection in two ways. First, we suggest the training procedure with over-sampling methods to obtain diverse training data sets. Second, we use some variants of accuracy and diversity measures that focus on fraud class. We also dynamically calculate the diversity measure-Forward Addition and Backward Elimination. In our experiments, Neural Networks, Decision Trees and Logit Regressions are the base models as the ensemble members and the performance of homogeneous ensembles are compared with that of heterogeneous ensembles. The experimental results show that the reduced size ensemble is as accurate on average over the data-sets tested as the non-pruned version, which provides benefits in terms of its application efficiency and reduced complexity of the ensemble.

• Journal of KIISE:Software and Applications
• /
• v.31 no.12
• /
• pp.1637-1648
• /
• 2004

When digital libraries are developed by the traditional client/sever system using a single agent on the distributed environment, several problems occur. First, as the search method is one dimensional, the search results have little relationship to each other. Second, the results do not reflect the user's preference. Third, whenever a client connects to the server, users have to receive the certification. Therefore, the retrieval of documents is less efficient causing dissatisfaction with the system. I propose a new platform of mobile multiagents for a personal digital library to overcome these problems. To develop this new platform I combine the existing DECAF multiagents platform with the Voyager mobile ORB and propose a new negotiation algorithm and scheduling algorithm. Although there has been some research for a personal digital library, I believe there have been few studies on their integration and systemization. For searches of related information, the proposed platform could increase the relationship of search results by subdividing the related documents, which are classified by a supervised neural network. For the user's preference, as some modular clients are applied to a neural network, the search results are optimized. By combining a mobile and multiagents platform a new mobile, multiagents platform is developed in order to decrease a network burden. Furthermore, a new negotiation algorithm and a scheduling algorithm are activated for the effectiveness of PDS. The results of the simulation demonstrate that as the number of servers and agents are increased, the search time for PDS decreases while the degree of the user's satisfaction is four times greater than with the C/S model.

• KIPS Transactions on Computer and Communication Systems
• /
• v.5 no.8
• /
• pp.181-188
• /
• 2016

Recently, as the amount of spatial information increases, an interest in the study of spatial information processing has been increased. Spatial database systems extended from the traditional relational database systems are difficult to handle large data sets because of the scalability. SpatialHadoop extended from Hadoop system has a low performance, because spatial computations in SpationHadoop require a lot of write operations of intermediate results to the disk, resulting in the performance degradation. In this paper, Spatial Computation Spark(SC-Spark) is proposed, which is an in-memory based distributed processing framework. SC-Spark is extended from Spark in order to efficiently perform the spatial operation for large-scale data. In addition, SC-Spark based on the GPGPU is developed to improve the performance of the SC-Spark. SC-Spark uses the advantage of the Spark holding intermediate results in the memory. And GPGPU-based SC-Spark can perform spatial operations in parallel using a plurality of processing elements of an GPU. To verify the proposed work, experiments on a single AMD system were performed using SC-Spark and GPGPU-based SC-Spark for Point-in-Polygon and spatial join operation. The experimental results showed that the performance of SC-Spark and GPGPU-based SC-Spark were up-to 8 times faster than SpatialHadoop.

• Journal of Dairy Science and Biotechnology
• /
• v.35 no.4
• /
• pp.255-261
• /
• 2017

The purpose of this study was to investigate the probiotic properties and antioxidant capacity of lactic acid bacteria isolated from Vietnamese feces and the Korean traditional food kimchi. Six isolated strains were identified as Lactobacillus sp. by 16S rRNA sequencing. All strains showed good resistance to low pH (1.5, 2.0, and 3.0) and 0.3% oxgall bile acids. Culture filtrates from the six strains showed various antioxidant effects, including DPPH, ABTS, reducing power, and metal chelating ($Fe^{2+}$) activities. Two of the six Lactobacillus strains showed potential probiotic activity. Heat resistance and adhesion assays were conducted by mixing the selected strains, Lactobacillus acidophilus V4, Lactobacillus plantarum V7, and Lactobacillus paracasei DK121 isolated from kimchi. The results showed that the heat resistance of these strains was similar to that of a commercial strain, L. plantarum LP. In addition, a mucin attachment assay using the mixture of selected strains (V4, V7, and DK121) showed high binding activity to the mucous layer. In conclusion, a mixture of V4, V7, and DK121 shows promising probiotic activity and may be useful for the development of health-related products.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.10 no.8
• /
• pp.1783-1788
• /
• 2009

We newly propose that we may reuse scrap powders ($Z_rO_2$+8 mol%-$Y_2O_3$) as $Y_2O_3$-stabilized zirconia (YSZ) sintered products through sintering process at 1550$^{circ}C$ for 2hrs. We also prepared the reference specimen from fresh $Z_rO_2$+30 mol%-$Y_2O_3$ powder mixture (celluar type with 1㎛-length). The reference sample showed a dense microstructure with grains of $\sim$10㎛ in diameter, while the sintered sample from scrap powder showed irregular grains of 1$\sim$30 ㎛ in diameter. Through XRD analysis, we confirmed that the reference sample has mixed phases of $Y_2O_3$(cubic), $Z_{r0.8}O_{1.9}$(cubic), and $Z_rO_2$(monoclinic), while the sintered YSZ sample from scrap powder has only tetragonal phase. Moreover, the sintered YSZ from scrap powder showed vickers hardness and apparent density more than 70 and 4.11 g/cc, which implies that it can be suitable for structural material application.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2000.02a
• /
• pp.78-78
• /
• 2000

다결정 실리콘-게르마늄 (poly-SiGe)은 태양전지 개발에 있어서 중요한 물질이다. 우리는 소량의 Ge(x=0.05)으로부터 다량의 Ge(x=0.67)을 함유한 수소화된 비정질 실리콘-게르마늄 (a-SiGe:H) 박막의 고상결정화 과정을 ESR (electron spin resonance)방법으로 조사해보았다. 먼저 PECVD 방법으로 Corning 1737 glass 위에 a-Si1-xGex:H 박막을 증착시켰다. 증착가스는 SiH4, GeH4 가스를 썼으며, 기판온도는 20$0^{\circ}C$, r.f. 전력은 3W, 증착시 가스압력은 0.6 Torr 정도이었다. 증착된 a-SiGe:H 박막은 $600^{\circ}C$ N2 분위기에서 다시 가열되어 고상결정화 되었고, 결정화 정도는 XRD (111) peak의 세기로부터 구해졌다. ESR 측정은 상온 x-band 영역에서 수행되었다. 측정된 ESR스팩트럼은 두 개의 Gaussian 함수로써 Si dangling-bond와 Ge dangling-bond 신호로 분리되었다. 가열 초기의 a-SiGe:H 박막 결함들의 스핀밀도의 증가는 수소 이탈에 기인하고, 또 고상결정화 과정에서 결정화된 정도와 Ge-db 스핀밀도의 변화는 서로 깊은 상관관계가 있음을 알 수 있었다. 특히 Ge 함유량이 큰 박막 (x=0.21, 0.67)에서 뿐만 아니라 소량의 Ge이 함유된 박막(x=0.05)에서도 Ge dangling-bond가 Si dangliong-bond 보다 고상결정화 과정에서 더 중요한 역할을 한다는 것을 알수 있었다. 또한 초기 열처리시 Si-H, Ge-H 결합에서 H의 이탈로 인하여 나타나는 Si-dangling bond, Ge-dangling bond 스핀밀도의 최대 증가 시간은 x 값에 의존하였는데 이러한 결과는 x값에 의존하는 Si-H, Ge-H 해리에너리지로 설명되어 질 수 있다. 층의 두께가 500 미만인 커패시터의 경우에 TiN과 Si3N4 의 계면에서 형성되는 슬릿형 공동(slit-like void)에 의해 커패시터의 유전특성이 파괴된다는 사실을 알게 되었으며, 이러한 슬릿형 공동은 제조 공정 중 재료에 따른 열팽창 계수와 탄성 계수 등의 차이에 의해 형성된 잔류응력 상태가 유전막을 기준으로 압축응력에서 인장 응력으로 바뀌는 분포에 기인하였다는 사실을 확인하였다.SiO2 막을 약화시켜 절연막의 두께가 두꺼워졌음에도 기존의 SiO2 절연막의 절연 파괴 전압 및 누설 전류오 비교되는 특성을 가졌다. 이중막을 구성하고 있는 안티퓨즈의 ON-저항이 단일막과 비교해 비슷한 것을 볼 수 잇는데, 그 이유는 TiO2에 포함된 Ti가 필라멘트에 포함되어 있어 필라멘트의 저항을 감소시켰기 때문으로 사료된다. 결국 이중막을 구성시 ON-저항 증가에 의한 속도 저하 요인은 없다고 할 수 있다. 5V의 절연파괴 시간을 측정한느 TDDB 테스트 결과 1.1$\times$103 year로 기대수치인 수십 년보다 높아 제안된 안티퓨즈의 신뢰성을 확보 할 수 있었다. 제안된 안티퓨즈의 이중 절연막의 두께는 250 이고 프로그래밍 전압은 9.0V이고, 약 65$\Omega$의 on 저항을 얻을수 있었다.보았다.다.다양한 기능을 가진 신소재 제조에 있다. 또한 경제적인 측면에서도 고부가 가치의 제품 개발에 따른 새로운 수요 창출과 수익률 향상, 기존의 기능성 안료를 나노(nano)화하여 나노 입자를 제조, 기존의 기능성 안료에 대한 비용 절감 효과등을 유도 할 수 있다. 역시 기술적인 측면에서도 특수소재 개발에 있어 최적의 나노 입자 제어기술 개발 및 나노입자를 기능성 소재로 사용하여 새로운 제품의 제조와 고압 기상 분사기술의 최적화에 의한 기능성 나노 입자 제조 기술을 확립하고 2차 오염 발생원인 유기계 항균제를 무기계 항균제로 대체할 수 있다. 이와 더불

• Korean Journal of Soil Science and Fertilizer
• /
• v.35 no.6
• /
• pp.344-351
• /
• 2002

Chemical speciation and transport of heavy metals in soil environment could be controlled by humic acid. In order to understand the mechanism on competitive adsorption of Cd and Cu on the surface of humic acid extracted from peat, the charge development of humic acid were investigate through a batch adsorption experiment with a series of different background electrolytes levels. The competitive adsorption of Cd and Cu to the humic acid were estimated according to the model using the proton binding constant obtained from the above batch test. The affinity of Cu to the carboxyl group on the humic acid was higher than that of Cd, but the affinity to the phenolic group was lower than to the carboxyl group. It seems that the amount of adsorbed Cd and Cu could be estimated using the proton binding constant obtained from a solution with single background ion. However, it is difficult to interpret the competitive adsorption of Cd and Cu with the constant for single background ion.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2010.02a
• /
• pp.406-406
• /
• 2010

일반적으로, 나노스케일의 MOS 소자에서는 게이트 절연체 두께가 감소함에 따라 tunneling effect의 증가로 인해 PID (plasma induced damage)로 인한 소자 특성 저하 현상을 감소하는 추세로 알려져 있다. 하지만 요즘 많이 사용되고 있는 high-k 게이트 절연체의 경우에는 오히려 더 많은 charge들이 trapping 되면서 PID가 오히려 더 심각해지는 현상이 나타나고 있다. 이러한 high-k 게이트 식각 시 현재는 주로 Hf-based wet etch나 dry etch가 사용되고 있지만 gate edge 영역에서 high-k 게이트 절연체의 undercut 현상이나 PID에 의한 소자특성 저하가 보고되고 있다. 본 연구에서는 이에 차세대 MOS 소자의 gate stack 구조중 issue화 되고 있는 metal gate 층과 gate dielectric 층의 식각공정에 각각 중성빔 식각과 중성빔 원자층 식각을 적용하여 전기적 손상 없이 원자레벨의 정확한 식각 조절을 해줄 수 있는 새로운 two step 식각 공정에 대한 연구를 진행하였다. 먼저 TiN metal gate 층의 식각을 위해 HBr과 $Cl_2$ 혼합가스를 사용한 중성빔 식각기술을 적용하여 100 eV 이하의 에너지 조건에서 하부층인 $HfO_2$와 거의 무한대의 식각 선택비를 얻었다. 하지만 100 eV 조건에서는 낮은 에너지에 의한 빔 스케터링으로 실제 패턴 식각시 etch foot이 발생되는 현상이 관찰되었으며, 이를 해결하기 위하여 먼저 높은 에너지로 식각을 진행하고 $HfO_2$와의 계면 근처에서 100 eV로 식각을 해주는 two step 방법을 사용하였다. 그 결과 anistropic 하고 하부층에 etch stop된 식각 형상을 관찰할 수 있었다. 다음으로 3.5nm의 매우 얇은 $HfO_2$ gate dielectric 층의 정확한 식각 깊이 조절을 위해 $BCl_3$와 Ar 가스를 이용한 중성빔 원자층 식각기술을 적용하여 $1.2\;{\AA}$/cycle의 단일막 식각 조건을 확립하고 약 30 cycle 공정시 3.5nm 두께의 $HfO_2$ 층이 완벽히 제거됨을 관찰할 수 있었다. 뿐만 아니라, vertical 한 식각 형상 및 향상된 표면 roughness를 transmission electron microscope(TEM)과 atomic force microscope (AFM)으로 관찰할 수 있었다. 이러한 중성빔 식각과 중성빔 원자층 식각기술이 결합된 새로운 gate recess 공정을 실제 MOSFET 소자에 적용하여 기존 식각 방법으로 제작된 소자 결과를 비교해 본 결과 gate leakage current가 약 one order 정도 개선되었음을 확인할 수 있었다.

• Korean Journal of Materials Research
• /
• v.3 no.3
• /
• pp.261-271
• /
• 1993

Functionally Gradient Materials(FGM) of 4.5Pb($Ni_{1/3}Nb_{2/3})O_3$-55PZT and PLZT(lO/70 /30, 11/60/40) were prepared. Its dielectric and piezoelectric strain properties were investigated. The FGM were pressed into A/B/ A configuration using two kinds of films, one layer(A) was eliminated from FGM by polishing after sintering at $l250^{\circ}C$, 2 hrs. The acrylic binder system was successfully applied for crack free film through doctor blade method. The thickness of gradent layer in FGM was about 30${\mu}$m. Dielectric properties of FGM show the average value of each side layer. The strain-electric field characteristics of FGM were significantly improved comparison with the other single compositions. The prepared FGM piezoelectric actuator shows about 3${\mu}$m/IOOV displacement.

• Korean Journal of Remote Sensing
• /
• v.23 no.4
• /
• pp.257-272
• /
• 2007

This paper analyzed relationships between various features from SAR data with multiple acquisition dates and mode (frequency, polarization and incidence angles), and land-cover classes. Two typical types of features were extracted by considering acquisition conditions of currently available SAR data. First, coherence, temporal variability and principal component transform-based features were extracted from multi-temporal and single mode SAR data. C-band ERS-1/2, ENVISAT ASAR and Radarsat-1, and L-band JERS-1 SAR data were used for those features and different characteristics of different SAR sensor data were discussed in terms of land-cover discrimination capability. Overall, tandem coherence showed the best discrimination capability among various features. Long-term coherence from C-band SAR data provided a useful information on the discrimination of urban areas from other classes. Paddy fields showed the highest temporal variability values in all SAR sensor data. Features from principal component transform contained particular information relevant to specific land-cover class. As features for multiple mode SAR data acquired at similar dates, polarization ratio and multi-channel variability were also considered. VH/VV polarization ratio was a useful feature for the discrimination of forest and dry fields in which the distributions of coherence and temporal variability were significantly overlapped. It would be expected that the case study results could be useful information on improvement of classification accuracy in land-cover classification with SAR data, provided that the main findings of this paper would be confirmed by extensive case studies based on multi-temporal SAR data with various modes and ground-based SAR experiments.