• Journal of Electrical Engineering and Technology
• /
• 제13권4호
• /
• pp.1724-1731
• /
• 2018

The conventional polynomial neural network (PNN) is a classical flexible neural structure and self-organizing network, however it is not free from the limitation of overfitting problem. In this study, we propose a space search-optimized polynomial neural network (ssPNN) structure to alleviate this problem. Ranking selection is realized by means of ranking selection-based performance index (RS_PI) which is combined with conventional performance index (PI) and coefficients based performance index (CPI) (viz. the sum of squared coefficient). Unlike the conventional PNN, L2-norm regularization method for estimating the polynomial coefficients is also used when designing the ssPNN. Furthermore, space search optimization (SSO) is exploited here to optimize the parameters of ssPNN (viz. the number of input variables, which variables will be selected as input variables, and the type of polynomial). Experimental results show that the proposed ranking selection-based polynomial neural network gives rise to better performance in comparison with the neuron fuzzy models reported in the literatures.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2004년도 Proceedings of ISRS 2004
• /
• pp.673-676
• /
• 2004

To manage and query moving objects efficiently in MMDBMS, a memory index structure should be used. The most popular index structure for storing trajectories of moving objects is 3DR-tree. The 3DR-tree also can be used for MMDBMS. However, the volume of data can exceed the capacity of physical memory since moving objects report their locations continuously. To accommodate new location reports, old trajectories should be migrated to disk or purged from memory. This paper focuses on migration policies of a main memory index structure. Migration policies consist of two steps: (i) node selection, (ii) node placement. The first step (node selection) selects nodes that should be migrated to disk. The criteria of selection are the performance of insertion or query. The second step (node placement) determines the order of nodes written to disk. This step can be thought as dynamic declustering policies.

• Journal of Information Technology Applications and Management
• /
• 제10권2호
• /
• pp.19-31
• /
• 2003

We propose an indexing technique and an index selection algorithm for optimal OLAP query processing in object database systems, Although there are many research results on the relational database systems for OLAP Query processing, few researches have been done on the object database systems. Since OLAP queries represent complex business logic on a huge data ware-house, object database systems supporting the OLAP queries should have higher performance. Proposed bitmap index structure is an extension of conventional bitmap indexes for adapting object databases and provides higher performance with lower space overhead. We also propose a linear time solution of the index selection problem that will be used in the OLAP query optimization process.

• Animal Bioscience
• /
• 제37권5호
• /
• pp.839-851
• /
• 2024

Objective: The main goal of our current study was to improve the growth curve of meat animals by decreasing the birth weight while achieving a finishing weight that is the same as that before selection but at younger age. Methods: Random regression model was developed to derive various selection indices to achieve desired gains in body weight at target time points throughout the fattening process. We considered absolute and proportional gains at specific ages (in weeks) and for various stages (i.e., early, middle, late) during the fattening process. Results: The point gain index was particularly easy to use because breeders can assign a specific age (in weeks) as a time point and model either the actual weight gain desired or a scaled percentage gain in body weight. Conclusion: The point gain index we developed can achieve the desired weight gain at any given postnatal week of the growing process and is an easy-to-use and practical option for improving the growth curve.

• 한국에너지공학회:학술대회논문집
• /
• 한국에너지공학회 1999년도 추계 학술발표회 논문집
• /
• pp.119-124
• /
• 1999

A selection methodology of liquid levelmeters, especially, level sensors in non-nuclear category, to be installed in nuclear energy facilities is developed using linguistic fuzzy approaches such as fully-linguistic and semi-linguistic methods. Depending on defuzzification techniques, the linguistic fuzzy methodology leads to either linguistic (exactly, fully-linguistic) or cardinal (i.e., semi-linguistic) evaluation. For the linguistic method, for each alternative, fuzzy preference index is converted to linguistic utility value by means of a similarity measure determining the degree of similarity between fuzzy index and linguistic ratings. For the cardinal method, the index is translated to cardinal overall utility value. According to these values, alternatives of interest are linguistically or numerically evaluated and a suitable alternative can be selected. Under given selection criteria, the suitable selections out of some liquid levelmeters for nuclear facilities are dealt with using the linguistic fuzzy methodology proposed. Then, linguistic fuzzy evaluation results are compared with qualitative result available in the literature. It is found that as to a suitable option the linguistic fuzzy selection is in agreement with the qualitative selection. Additionally, the comparative study shows that the fully-linguistic method using adequate scale system facilitates linguistic interpretation regarding evaluation results.

• Journal of information and communication convergence engineering
• /
• 제9권5호
• /
• pp.491-499
• /
• 2011

In this paper, we offer a new strategic portfolio model for national IT R&D project selection in Korea. A risk and return (R-R) portfolio model was developed using an objectively quantified index on the two axes of risk and return, in order to select a strategic project and allocate resources in compliance with a national IT R&D strategy. We strategize using the R-R portfolio model to solve the non-strategy and subjectivity problems of the existing national R&D project selection model. We also use the quantified evaluation index of the IT technology road map (TRM) and the technical level reports (TLR) for the subjectivity of project selection, and try to discover the weights using the analytic hierarchy process (AHP). In addition, we intend to maximize the chance for a successful national IT R&D project, by selecting a strategic portfolio project and balancing the allocation of resources effectively and objectively.

• 대한토목학회논문집
• /
• 제37권5호
• /
• pp.855-862
• /
• 2017

대한민국은 지난 반세기간 성장위주의 정책을 통하여서 경제적, 문화적, 사회적으로 급속한 발전을 이루었다. 하지만 거점개발방식으로 인하여서 비대한 도시와 낙후된 지역이 생겼다. 이에 따라서 국가에서는 많은 낙후지역 개발정책들을 펼쳐왔으나 많은 문제점들을 가졌다. 그 이유는 소관 부처가 많고 사업의 종류가 다양함으로 인해 나타난다. 또한 개발 우선순위가 객관적인 지표에 근거하지 않고 각 지자체에 분배 식으로 사업이 수행되는 형태이기 때문이다. 기존의 연구된 지역에 대한 지표연구는 도 또는 시, 군등 대단위에서만 분석되어있고 소단위 지역의 사업의 순서를 선정하는데 어려움을 가진다. 이에 따라서 지역에 대한 지표선정을 통해 소규모 낙후지역의 효율적인 개발을 위한 의사결정 지원 시스템로 사용하고자 한다.

• International Journal of Aeronautical and Space Sciences
• /
• 제16권3호
• /
• pp.437-444
• /
• 2015

Satellite selection and exclusion techniques have been applied to the global navigation satellite system (GNSS) with the aim of achieving a balance between navigational performance and computational efficiency. Conventional approaches to satellite selection based on the best dilution of precision (DOP) are excessively computational and complicated. This paper proposes a new method that applies a geometric sensitivity index of individual GNSS satellites. The sensitivity index is derived using the inner product of the line of sight (LOS) vector of each satellite. First, the LOS vector is computed, which accounts for the geometry between the satellite and user positions. Second, the inner product of each pair of LOS vectors is calculated, which indicates the proximities of the satellites to one another. The proximity can be determined according to the sensitivity of each satellite. A post-processing test was conducted to verify the reliability of the proposed method. The proposed index and the results of a conventional approach that measures the dilution of precision (DOP) were compared. The test results demonstrate that the proposed index produces results that are within 96% of those of the conventional approach and reduces the computational burden. This index can be utilized to estimate the sensitivity of individual satellites, obtaining a navigation solution. Therefore, the proposed index applies to satellite selection and exclusion as well as to the sensitivity analyses of multiple GNSS applications.

• 농업과학연구
• /
• 제20권1호
• /
• pp.48-55
• /
• 1993

본 연구(硏究)는 1986년부터 1990년까지 충청남도(忠淸南道) 도립종축장(道立種畜場)에서 사육된 Landrace, Large Yorkshire 및 Duroc종의 후보종 모돈 115두의 이유후 형질(形質)에 대한 능력검정(能力檢定) 자료를 근거로 최소자승법을 이용하여 품종(品種), 산차(産差), 분만년도(分娩年度) 및 분만월(分娩月)이 일당증체량(日當增體量), 사료요구율, 등지방층두께, 90kg도달일령 및 선발지수에 미치는 효과(效果)를 추정(推定)하였던 바 그 결과(結果)를 요약(要約)하면 다음과 같다. 1. 이유후형질(形質)의 전체평균은 일당증체량(日當增體量), 사료요구율, 등지방층두께, 90kg도달일령 및 선발지수(選拔指數)에서 각각 $922.25{\pm}17.25$g, $2.85{\pm}0.04$, $2.03{\pm}0.04$cm, $135.44{\pm}1.38$일 및 $168.33{\pm}2.42$이었다. 2. 품종의 효과(效果)에 있어서는 일당증체량(日當增體量), 90kg도달일령 및 선발지수(選拔指數)에서 고도의 유의차(有意差)(P<0.01)가 인정되었으며, 등지방두께에 있어서도 유의차(P<0.05)가 나타났다. 품종별로는 Duroc종이 일당증체량(日當增體量), 90Kg도달일령및 선발지수(選拔指數)에서는 각각 977.23g, 132.47일 및 172.35로 가장 우수하였고, 등지방층두께는 Landrace종이 1.95cm로 가장 얇었다. 3. 산차(産差)의 효과(效果)에 있어서 일당증체량(日當增體量)은 고도의 유의차(有意差)(P<0.01)가 인정되었으며, 등지방층두께 및 선발지수(選拔指數)에서도 유의차(有意差)(P<0.05)가 인정되었고, 산차(産差)로는 일당증체량(日當增體量) 및 선발지수(選拔指數)가 3산차(産差)와 4산차(産差)에서 각각 974.92g, 177.61, 959.48g 및 1774.84로 가장 우수하였다. 4. 분만년도의 효과(效果)에 있어서는 일당증체량(日當增體量), 사료요구율 및 선발지수(選拔指數)에서 고도의 유의차(P<0.01)가 인정되었으며, 분만월에 있어서도 일당증체량(日當增體量)에서 고도의 有意差(P<0.01)가 인정되었고, 등지방층두께 및 선발지수(選拔指數)에서는 유의차(有意差)(P<0.05)가 있었으며, 분만월별로는 3월분만이 일당증체량 및 선발지수(選拔指數)에서 각각 968.22g, 174.54로 가장 우수하였다.

• 정보과학회 컴퓨팅의 실제 논문지
• /
• 제21권1호
• /
• pp.76-81
• /
• 2015

최근 대용량 데이터의 분석을 위한 top-k 질의처리 알고리즘에 대한 관심이 고조되고 있다. 그러나 기존 top-k 질의처리 알고리즘은 효율적인 인덱스 구조를 제공하지 않기 때문에, 높은 탐색 비용을 야기하며, 아울러 다양한 질의 유형을 지원하지 못하는 문제점이 존재한다. 이러한 문제점을 해결하기 위해, 본 논문에서는 그리드 인덱스 기반 뷰 선택 기법을 이용한 top-k 질의처리 알고리즘을 제안한다. 제안하는 기법은 그리드 인덱스 기반의 뷰 선택 기법을 통해 주어진 질의 영역에 대하여 최소한의 그리드 셀만을 탐색함으로써 질의처리 시간을 감소시킨다. 마지막으로, 성능 평가를 통해 제안하는 top-k 질의처리 알고리즘이 기존 알고리즘에 비해 질의처리 시간 및 질의 결과 정확도 측면에서 우수함을 나타낸다.