• Journal of Applied Reliability
• /
• v.16 no.2
• /
• pp.90-97
• /
• 2016

Purpose: Majority of bridges and roads in Gangwon Province have been carrying loads associated with heavy materials such as rocks, mining products, and cement. This location-specific live loads have contributed to the present situation of overloading, compared to other provinces in Korea. However, the bridges in Gangwon province are designed by national bridge design specification, without considering the location-specific live load characteristics. Therefore, this study focuses on the real traffic data accumulated on regional weighing station to verify the live load characteristics, including actual live load gross vehicle weight, axle weight axle spacings, and number of trucks. Methods: In order to take into account the location specific live load, a governmental weigh station (38th national highway Miro) have been selected and the passing truck data are processed. Based on the truck survey, trucks are categorized into 3 different shapes, and each shape has been idealized into normal distribution. Then, the resulting survey data are processed to predict the target maximum live load values, including the axle loads and gross vehicle weights in 75 years service life span. Results: The results are compared to the nationally used DB-24 live loads, and the results show that nationally recognized DB-24 live load does not sufficiently represent real traffic in mountaineous region in Gangwon province. Conclusion: The comparison results in the recommendation of location-specific live load that should be taken into account for bridge design and evaluation.

• Journal of Korean Institute of Industrial Engineers
• /
• v.31 no.3
• /
• pp.257-264
• /
• 2005

This paper suggests a heuristic algorithm for the clustering problem. Clustering involves grouping similar objects into a cluster. Clustering is used in a wide variety of fields including data mining, marketing, and biology. Until now there are a lot of approaches using Self-Organizing Feature Maps(SOFMs). But they have problems with a small output-layer nodes and initial weight. For example, one of them is a one-dimension map of k output-layer nodes, if they want to make k clusters. This approach has problems to classify elaboratively. This paper suggests one-dimensional output-layer nodes in SOFMs. The number of output-layer nodes is more than those of clusters intended to find and the order of output-layer nodes is ascending in the sum of the output-layer node's weight. We can find input data in SOFMs output node and classify input data in output nodes using Euclidean distance. We use the well known IRIS data as an experimental data. Unsupervised clustering of IRIS data typically results in 15 - 17 clustering error. However, the proposed algorithm has only six clustering errors.

• ETRI Journal
• /
• v.27 no.4
• /
• pp.385-393
• /
• 2005

This paper presents an application of digital signal processing to data acquired by the radio imaging method (RIM) that was adopted to measure moisture distribution inside the human body. RIM was originally developed for the mining industry; we are applying the method to a biomedical measurement because of its simplicity, economy, and safety. When a two-dimensional image was constructed from the measured data, the method provided insufficient resolution because the wavelength of the measurement medium, a weak electromagnetic wave in a VHF band, was longer than human tissues. We built and measured a phantom, a model simulating the human body, consisting of two water tanks representing large internal organs. A digital equalizer was applied to the measured values as a weight function, and images were reconstructed that corresponded to the original shape of the two water tanks. As a result, a two-dimensional image containing two individual peaks corresponding to the original two small water tanks was constructed. The result suggests the method was applicable to biomedical measurement by the assistance of digital signal processing. This technique may be applicable to home-based medical care and other situations in which safety, simplicity, and economy are important.

• ETRI Journal
• /
• v.39 no.5
• /
• pp.621-631
• /
• 2017

Abnormal samples are usually difficult to obtain in production systems, resulting in imbalanced training sample sets. Namely, the number of positive samples is far less than the number of negative samples. Traditional Support Vector Machine (SVM)-based anomaly detection algorithms perform poorly for highly imbalanced datasets: the learned classification hyperplane skews toward the positive samples, resulting in a high false-negative rate. This article proposes a new imbalanced SVM (termed ImSVM)-based anomaly detection algorithm, which assigns a different weight for each positive support vector in the decision function. ImSVM adjusts the learned classification hyperplane to make the decision function achieve a maximum GMean measure value on the dataset. The above problem is converted into an unconstrained optimization problem to search the optimal weight vector. Experiments are carried out on both Cloud datasets and Knowledge Discovery and Data Mining datasets to evaluate ImSVM. Highly imbalanced training sample sets are constructed. The experimental results show that ImSVM outperforms over-sampling techniques and several existing imbalanced SVM-based techniques.

• Journal of the Korea Society of Computer and Information
• /
• v.18 no.1
• /
• pp.177-183
• /
• 2013

In this paper, Cohen's kappa and weighted kappa are applied to measuring classification accuracy when performing classification in data minig. Cohen's kappa compensates for classifications that may be due to chance and is used for the data with nominal or ordinal scales. Especially, for the ordinal data, weighted kappa which measures the classification accuracy by quantifying the classification errors as weights is used. We used two weights (linear weight, quadratic weight) for calculations of weighted kappa. Also for the calculation and comparison of kappa and weighted kappa we used a real data set, fat-liver data.

• Advances in concrete construction
• /
• v.4 no.3
• /
• pp.221-230
• /
• 2016

This paper presents a preliminary study on the influence of laterite soil replacing conventional fine aggregates on the strength properties of GGBS-blended-concrete. For this purpose, GGBS-blended-concrete samples with 40% GGBS, 60% Portland cement (PC), and locally available laterite soil was used. Laterite soils at 0, 25, 50 and 75% by weight were used in trails to replace the conventional fine aggregates. A control mix using only PC, river sand, course aggregates and water served as bench mark in comparing the performance of the composite concrete mix. Test blocks including 60 cubes for compression test; 20 cylinders for split tensile test; and 20 beams for flexural strength test were prepared in the laboratory. Results showed decreasing trends in strength parameters with increasing laterite content in GGBS-blended-concrete. 25% and 50% laterite replacement showed convincing strength (with small decrease) after 28 day curing, which is about 87-90% and 72-85% respectively in comparison to that achieved by the control mix.

• Proceedings of the Korean Society of Soil and Groundwater Environment Conference
• /
• 1996.12a
• /
• pp.97-101
• /
• 1996

Municipal and sanitary landfills can pose environmental problems due to leachate, landfill gas md unstable geotechnical properties. Most governmental bodies delay the correction of landfill problems or landfill replacement until a crises stage is reached. The replacement of a landfill is often made difficult due to costly regulatory controls, public opposition to siting and the high cost of closure for the previous landfill unit. Solutions to extending landfill life and capacity Involve waste minimization by recycling, refuse compaction and waste-to-energy incineration. Incineration can reduce the volume of refuse by 50-95%. The largest installed bases of municipal waste Incinerators are located in Japan and the U.S. The volume of waste contained in a landfill can be estimated by load count tabulations, weight-and-volume measurements or a material balance analysis based on the trash profile of user categories. for an existing landfill, core samples may be collected and analyzed for use in a material balance analysis. Newly generated refuse contains approximately 50% of the heating value of coal. However, landfill properties vary significantly due to the waste profile of the contributors and biodegradation due to time and weathering. The volume of the Nanji-do landfill

• Proceedings of the Korean Information Science Society Conference
• /
• 2001.04b
• /
• pp.244-246
• /
• 2001

최근 군집화와 분류기법이 데이터 마이닝에 중요한 도구로 많은 응용분야에 사용되고 있다. 따라서 이러한 기법을 이용하는데 있어서 각각의 속성의 중요도가 달라 중요하지 않은 속성에 의해 중요한 속성이 왜곡되거나 때로는 마이닝의 결과가 잘못되는 결과를 얻을 수 있으며, 또한 전체 데이터를 사용할 경우 마이닝 과정을 저하시키는 문제로 속성 가중치과 속성선택에 과한 연구가 중요한 연구의 대상이 되고 있다. 최근 연구되고 있는 알고리즘들은 사용자의 의도와는 상관없이 데이터간의 관계에만 의존하여 가중치를 설정하므로 사용자가 마이닝 결과를 쉽게 이해하고 분석할 수 없는 문제점을 안고 있다. 본 논문에서는 클래스 정보가 있는 데이터뿐 아니라 클래스 정보가 없는 데이터를 분석할 경우 사용자의 의도에 따라 학습할 수 있도록 각 가중치를 부여하는 속성가중치 알고리즘을 제안한다. 또한 사용자가 의도한 정보를 이용하여 속성간의 가장 최적화 된 가중치를 찾아주며, Cramer's $V^2$함수를 적합도 함수로 하는 유전자 알고리즘을 사용한다. 알고리즘의 타당성을 검증하기 위해 전자상거래상의 실험 데이터와 몇 가지 벤치마크 데이터를 이용하여 본 논문의 타당성을 보인다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2017.11a
• /
• pp.395-398
• /
• 2017

We propose a web crawler system with keyword extraction function in this paper. Researches on the keyword extraction in existing text mining are mostly based on databases which have already been grabbed by documents or corpora, but the purpose of this paper is to establish a real-time keyword extraction system which can extract the keywords of the corresponding text and store them into the database together while grasping the text of the web page. In this paper, we design and implement a crawler combining RAKE keyword extraction algorithm. It can extract keywords from the corresponding content while grasping the content of web page. As a result, the performance of the RAKE algorithm is improved by increasing the weight of the important features (such as the noun appearing in the title). The experimental results show that this method is superior to the existing method and it can extract keywords satisfactorily.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2001.05a
• /
• pp.8-11
• /
• 2001

연관규칙 마이닝은 트랜잭션 데이터를 이루고 있는 항목간의 잠재적인 의존관계를 발견하는 데이터 마이닝의 한 분야이다. 정량 연관규칙이란 부류적 속성과 정량적 속성을 모두 포함한 연관규칙이다. 정량 연관규칙 마아닝을 위한 퍼지 기술의 응용, 정량 연관규칙 마이닝을 위한 일반화된 연관규칙 마이닝, 사용자의 관심도를 반영한 중요도 가중치가 있는 연관규칙 마이닝 등에 대한 연구가 이루어져 왔다. 이 논문에서는 중요도 가중치가 있는 일반화된 퍼지 정량 연관규칙 마이닝의 새로운 방법을 제안한다. 이 방법은 부류적 속성의 퍼지 개념 계층과 정량적 속성의 퍼지 언어항 일반화 계층을 일반화된 추출하기 위해 이용한다. 이것은 속성들의 수준별 일반화 계층과 속성의 중요도 가중치를 이용함으로써 사용자가 보다 융통성 있는 연관규칙을 마이닝할 수 있게 해준다.