• Journal of Intelligence and Information Systems
• /
• v.7 no.2
• /
• pp.19-33
• /
• 2001

Link weight analysis approach is suggested as a heuristic for selection of input nodes in artificial neural network for bankruptcy prediction. That is to analyze each input node\\\\`s link weight-absolute value of link weight between an input node and a hidden node in a well-trained neural network model. Prediction accuracy of three methods in this approach, -weak-linked-neurons elimination method, strong-linked-neurons selection method and integrated link weight model-is compared with that of decision tree and multivariate discrimination analysis. In result, the methods suggested in this study show higher accuracy than decision tree and multivariate discrimination analysis. Especially an integrated model has much higher accuracy than any individual models.

• Nuclear Engineering and Technology
• /
• v.30 no.1
• /
• pp.58-74
• /
• 1998

A PC window-based computer code, CONPAS (CONtainment Performance Analysis System), has been developed to integrate the numerical, graphical, and results-operation aspects of Level 2 probabilistic safety assessments (PSA) for nuclear power plants automatically. As a main logic for accident progression analysis, it employs a concept of the small containment phenomenological event tree (CPET) helpful to trace out visually individual accident progressions and of the detailed supporting event tree (DSET) for its detailed quantification. For the integrated analysis of Level 2 PSA, the code utilizes five distinct, but closely related modules. Its computational feasibility to real PSAs has been assessed through an application to the UCN 3&4 full scope Level 2 PSA. Compared with other existing computer codes for Level 2 PSA, the CONPAS code provides several advanced features: (1) systematic uncertainty analysis / importance analysis / sensitivity analysis, (2) table / graphical display & print, (3) employment of the recent Level 2 PSA technologies, and (4) highly effective user interface. The main purpose of this paper is to introduce the key features of CONPAS code and results of its feasibility study.

• Journal of Forest and Environmental Science
• /
• v.34 no.2
• /
• pp.119-125
• /
• 2018

The tropical forests in developing countries are faced with the problem of illegal exploitation of trees. However, dearth of empirical means of expressing the dimensions, structure, quality and quantity of a removed tree has imped conviction of offenders. This study aimed at developing a model that can effectively estimate individual tree basal area (BA) from stump diameter (Ds) for Tectona grandis stands in Omo Forest Reserve, Nigeria, for timber valuation in case of illegal felling. Thirty-six $25m{\times}25m$ temporary sample plots (TSPs) were laid randomly in six age strata; 26, 23, 22, 16, 14, and 12 years specifically. BA, Ds and diameter at breast height were measured in all living T. grandis trees within the 36 TSPs. Least square method was used to convert the counted stumps into harvested stem cross-sectional areas. Six basal area models were fitted and evaluated. The BA-Ds relationship was best described by power model which gave least values of Root mean square error (0.0048), prediction error sum of squares (0.0325) and Akaike information criterion (-15391) with a high adjusted coefficient of determination (0.921). This study revealed that basal area estimation was realistic even when the only information available was stump diameter. The power model was validated using independent data obtained from additional plots and was found to be appropriate for estimating the basal area of Tectona grandis stands in Omo Forest Reserve, Nigeria.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.57 no.2
• /
• pp.155-163
• /
• 2008

This paper presents a novel framework for analysis of power system wide-area blackout based on so called fault cascading scenarios. For a given power system operating state, "triggering" faults or a "seed faults" are chosen based on the probabilities estimated from the hazard rates. The fault probabilities reflect both the load and the weather conditions. Effects of hidden failures in protection systems are also reflected in establishing the fault propagation scenarios since they are one of the major causes for the wide-area blackouts. A tree type data structure called a PS-BEST(Power System Blackout Event Scenario Tree) is proposed for construction of the fault cascading scenarios, in which nodes represent various power system operating states and the arcs are the events causing transitions between the states. Arcs can be either probabilistic or deterministic. For a given initial fault, the total probability of leading to wide-area blackout is estimated by aggregating the individual probability of each fault sequence route leading to wide-area blackout. A case study is performed on the IEEE RTS-79(24 bus) system based on the fault data presented by the North American Electrical Reliability Council(NERC). Test results demonstrate the potentials and the effectiveness of the proposed technique for the future wide-area blackout analysis.

• Journal of Korea Water Resources Association
• /
• v.48 no.11
• /
• pp.945-955
• /
• 2015

The technology of the SWG (Smart Water Grid) as one of most important national projects results in significant assignment that is closely associated with systematic management and effective operation. The individual component technics are required to establish directory and classification for the purpose of effectively managing their information related to research and development (R&D). The national science technology (S&T) standard classification tree which results in the representative example has been established with an intention to manage R&D information, human resource, and budget. It has been also revised every five years and then used in the various fields related to the evaluation, administration, and prediction of the national R&D projects. In addition, the standard classification system for R&D projects has been widely used in the UNESCO (United Nations Educational, Scientific and Cultural Organization) and EU (European Union) since the Frascati Manual was established in the Organization for Economic Cooperation and Development (OECD). Therefore, it is necessary for SWG techniques to develop the standard S&T classification tree for research management and evaluation. For this, it is essential to draw the core techniques for the SWG, which are incorporated with IT (Information Technology), NT (Nano Technology), and BT (Biology Technology).

• Proceedings of the Korean Society of Surveying, Geodesy, Photogrammetry, and Cartography Conference
• /
• 2004.11a
• /
• pp.361-366
• /
• 2004

Geographic Information has been used widely for landuse and management, city plan, and environment and disaster management, etc., But geographic information has been built for individual cases using various methods. Therefore, the discordancy in data, double investment, confusion of use and difficulty of decision supporting system have been occurred. In order to solve these problems, national government is need to framework database. This framework database was enacted for building and use of National Geographic Information System and focused on basic plan of the second national geographic information system. Also, the framework database was selected of eight fields by NGIS laws and 19 detailed items through meeting of framework committee since 2002. In this research, The 19 detailed items( road, railroad, coastline, surveying control point etc.,) of framework database consider a Priority order, In the result of this research, the framework database is obtain to a priority order for building and the national government will carry effectively out a budget for the framework database building. Each of 19 detailed items is grouping into using the priority order of the framework database by AHP analysis method and verified items by decision tree analysis method. The one of the highest priority order items is a road, which is important for building, continuous renovation, and maintain management for use.

• Nuclear Engineering and Technology
• /
• v.50 no.5
• /
• pp.745-750
• /
• 2018

New exercise technology such as the virtual reality (VR)-based exercise system is required to meet soaring demand for target participants in exercises and to alleviate the difficulties in personnel mobilization through an alternative approach to the exercise system. In a previous study, event tree methodologies were introduced in setting up an exercise scenario of a VR-based radiological exercise system. In the scenario, the locations at which major events occur are rephrased as nodes, routes as paths, and public response actions as protective actions or contents of an exercise at individual locations. In the study, a model for estimating effective doses to the participants is proposed to evaluate the exercise system, using the effective dose rates at particular times and locations derived from a computer program. The effective dose received by a student when she/he follows a successful route is about a half of the dose received when she/he does not follow the exercise guide directions. In addition, elapsed time to finish an exercise when following a successful route is less than one-third of the time spent to finish an exercise when following the guide's directions.

• Journal of Korean Society of Water and Wastewater
• /
• v.35 no.4
• /
• pp.259-275
• /
• 2021

This study explored the usefulness and implications of the Bayesian hyperparameter optimization in developing species distribution models (SDMs). A variety of machine learning (ML) algorithms, namely, support vector machine (SVM), random forest (RF), boosted regression tree (BRT), XGBoost (XGB), and Multilayer perceptron (MLP) were used for predicting the occurrence of four benthic macroinvertebrate species. The Bayesian optimization method successfully tuned model hyperparameters, with all ML models resulting an area under the curve (AUC) > 0.7. Also, hyperparameter search ranges that generally clustered around the optimal values suggest the efficiency of the Bayesian optimization in finding optimal sets of hyperparameters. Tree based ensemble algorithms (BRT, RF, and XGB) tended to show higher performances than SVM and MLP. Important hyperparameters and optimal values differed by species and ML model, indicating the necessity of hyperparameter tuning for improving individual model performances. The optimization results demonstrate that for all macroinvertebrate species SVM and RF required fewer numbers of trials until obtaining optimal hyperparameter sets, leading to reduced computational cost compared to other ML algorithms. The results of this study suggest that the Bayesian optimization is an efficient method for hyperparameter optimization of machine learning algorithms.

• Journal of Microbiology and Biotechnology
• /
• v.29 no.5
• /
• pp.731-738
• /
• 2019

Endophytic fungi are an important component of plant microbiota, and have the excellent capacity for producing a broad variety of bioactive metabolites. These bioactive metabolites not only affect the survival of the host plant, but also provide valuable lead compounds for novel drug discovery. In this study, forty-two endophytic filamentous fungi were isolated from Ficus elastica leaves, and further identified as seven individual taxa by ITS-rDNA sequencing. The antimicrobial activity of these endophytic fungi was evaluated against five pathogenic microorganisms. Two strains, Fes1711 (Penicillium funiculosum) and Fes1712 (Trichoderma harzianum), displayed broad-spectrum bioactivities. Our following study emphasizes the isolation, identification and bioactivity testing of chemical metabolites produced by T. harzianum Fes1712. Two new isocoumarin derivatives (1 and 2), together with three known compounds (3-5) were isolated, and their structures were elucidated using NMR and MS. Compounds 1 and 2 exhibited inhibitory activity against Escherichia coli. Our findings reveal that endophytic fungi from the rubber tree F. elastica leaves exhibit unique characteristics and are potential producers of novel natural bioactive products.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.3
• /
• pp.85-90
• /
• 2021

In this paper, fall detection using individual wearable devices for older people is considered. To design a low-cost wearable device for reliable fall detection, we present a comprehensive analysis of two representative models. One is a machine learning model composed of a decision tree, random forest, and Support Vector Machine(SVM). The other is a deep learning model relying on a one-dimensional(1D) Convolutional Neural Network(CNN). By considering data segmentation, preprocessing, and feature extraction methods applied to the input data, we also evaluate the considered models' validity. Simulation results verify the efficacy of the deep learning model showing improved overall performance.