• 한국근적외분광분석학회:학술대회논문집
• /
• 한국근적외분광분석학회 2001년도 NIR-2001
• /
• pp.1244-1244
• /
• 2001

In linear discriminant analysis there are two important properties concerning the effectiveness of discriminant function modeling. The first is the separability of the discriminant function for different classes. The separability reaches its optimum by maximizing the ratio of between-class to within-class variance. The second is the stability of the discriminant function against noises present in the measurement variables. One can optimize the stability by exploring the discriminant variates in a principal variation subspace, i. e., the directions that account for a majority of the total variation of the data. An unstable discriminant function will exhibit inflated variance in the prediction of future unclassified objects, exposed to a significantly increased risk of erroneous prediction. Therefore, an ideal discriminant function should not only separate different classes with a minimum misclassification rate for the training set, but also possess a good stability such that the prediction variance for unclassified objects can be as small as possible. In other words, an optimal classifier should find a balance between the separability and the stability. This is of special significance for multivariate spectroscopy-based classification where multicollinearity always leads to discriminant directions located in low-spread subspaces. A new regularized discriminant analysis technique, the principal discriminant variate (PDV) method, has been developed for handling effectively multicollinear data commonly encountered in multivariate spectroscopy-based classification. The motivation behind this method is to seek a sequence of discriminant directions that not only optimize the separability between different classes, but also account for a maximized variation present in the data. Three different formulations for the PDV methods are suggested, and an effective computing procedure is proposed for a PDV method. Near-infrared (NIR) spectra of blood plasma samples from mastitic and healthy cows have been used to evaluate the behavior of the PDV method in comparison with principal component analysis (PCA), discriminant partial least squares (DPLS), soft independent modeling of class analogies (SIMCA) and Fisher linear discriminant analysis (FLDA). Results obtained demonstrate that the PDV method exhibits improved stability in prediction without significant loss of separability. The NIR spectra of blood plasma samples from mastitic and healthy cows are clearly discriminated between by the PDV method. Moreover, the proposed method provides superior performance to PCA, DPLS, SIMCA and FLDA, indicating that PDV is a promising tool in discriminant analysis of spectra-characterized samples with only small compositional difference, thereby providing a useful means for spectroscopy-based clinic applications.

• 한국근적외분광분석학회:학술대회논문집
• /
• 한국근적외분광분석학회 2001년도 NIR-2001
• /
• pp.1042-1042
• /
• 2001

In linear discriminant analysis there are two important properties concerning the effectiveness of discriminant function modeling. The first is the separability of the discriminant function for different classes. The separability reaches its optimum by maximizing the ratio of between-class to within-class variance. The second is the stability of the discriminant function against noises present in the measurement variables. One can optimize the stability by exploring the discriminant variates in a principal variation subspace, i. e., the directions that account for a majority of the total variation of the data. An unstable discriminant function will exhibit inflated variance in the prediction of future unclassified objects, exposed to a significantly increased risk of erroneous prediction. Therefore, an ideal discriminant function should not only separate different classes with a minimum misclassification rate for the training set, but also possess a good stability such that the prediction variance for unclassified objects can be as small as possible. In other words, an optimal classifier should find a balance between the separability and the stability. This is of special significance for multivariate spectroscopy-based classification where multicollinearity always leads to discriminant directions located in low-spread subspaces. A new regularized discriminant analysis technique, the principal discriminant variate (PDV) method, has been developed for handling effectively multicollinear data commonly encountered in multivariate spectroscopy-based classification. The motivation behind this method is to seek a sequence of discriminant directions that not only optimize the separability between different classes, but also account for a maximized variation present in the data. Three different formulations for the PDV methods are suggested, and an effective computing procedure is proposed for a PDV method. Near-infrared (NIR) spectra of blood plasma samples from daily monitoring of two Japanese cows have been used to evaluate the behavior of the PDV method in comparison with principal component analysis (PCA), discriminant partial least squares (DPLS), soft independent modeling of class analogies (SIMCA) and Fisher linear discriminant analysis (FLDA). Results obtained demonstrate that the PDV method exhibits improved stability in prediction without significant loss of separability. The NIR spectra of blood plasma samples from two cows are clearly discriminated between by the PDV method. Moreover, the proposed method provides superior performance to PCA, DPLS, SIMCA md FLDA, indicating that PDV is a promising tool in discriminant analysis of spectra-characterized samples with only small compositional difference.

• 인터넷정보학회논문지
• /
• 제5권1호
• /
• pp.67-84
• /
• 2004

실시간 데이터베이스 시스템의 트랜잭션은 마감시간과 같은 시간 제약조건을 갖는다. 따라서 트랜잭션은 마감시간 내에 완료되어야 하며 동시에 데이터 일관성 제약조건을 만족해야 한다. 빠른 평균 응답시간 제공을 목표로 하는 기존 데이터베이스 시스템과 달리 실시간 데이터베이스 시스템은 트랜잭션의 마감시간 초과 비율과 마감시간을 초과한 트랜잭션에 의해 발생되는 비용에 의해 평가된다. 따라서 실시간 트랜잭션은 공정성과 시스템 처리율의 저하를 발생시킨다 하더라도 트랜잭션의 중요도 및 마감시간에 따라 스케줄링 되어야 하고, 항상 높은 우선 순위 트랜잭션의 선행 처리가 보장되어야 한다. 본 논문에서는 기존 실시간 스케줄링 알고리즘들이 갖는 문제점을 개선시킬 수 있는 새로운 스케줄링 알고리즘(Multi-level EFDF)과 펌, 소프트 실시간 트랜잭션온 고려한 병행실행제어 알고리즘(2PL-FT)을 제안한다. 또한 성능평가를 통해 2PL-FT와 AVCC방법에 대한 트랜잭션의 재시작 비율 및 마감시간 초과 비율을 비교한다. 이 실험들을 통해 제안하는2PL-FT알고리즘이 기존에 제안된 방법들보다 더 우수함을 보인다.

• 한국지반공학회논문집
• /
• 제24권10호
• /
• pp.57-70
• /
• 2008

피에조콘 관입시험의 장점은 연속적인 데이터의 취득이 보장되며 결국 대상지반의 신뢰성 있는 분석이 가능하다는 점이다. 따라서 지난 수십년간 국내외에서 콘 관입시험결과로부터 흙분류를 수행하는 많은 연구가 진행되었으며 차트나 도표 등의 형태로 흙분류 방법들이 제안되었다. 그러나 대부분의 차트 또는 방법들은 한국을 제외한 세계 각국의 자료들을 바탕으로 제안되어 국내 지반의 적용성에 대한 검증이 이루어져야 한다. 뿐만 아니라 기존 방법들에서는 사용된 입력자료에 따라 흙분류 결과가 상이한 경우가 있어 적용과 판단에 어려움이 있다. 그러나 불행히도 이러한 차트 형태로 제안된 기존 도표의 경우 지역성 등이 반영되어 수정 또는 보완이 필요하나 수정에 어려움이 있거나 거의 불가능하다. 이에 본 연구에서는 국내 17개 현장에서 수행된 피에조콘 관입시험결과와 채취된 시료에 대한 주상도 및 흙분류결과를 바탕으로 클러스터링 기법과 뉴로-퍼지 이론을 이용한 흙분류 모델을 제안하였다. 제안된 모델을 검증하기 위해 모델 학습 시 사용되지 않는 새로운 피에조콘 관입시험 데이터에 대한 흙분류 결과를 실제 시추결과와 비교하였다. 또한 기존의 소프트컴퓨팅 모델과 Robertson 방법에 의한 흙분류 결과와 제안된 모델의 흙분류 결과를 비교하여 제안된 모델의 효율성을 검토하였다.

• Asia Marketing Journal
• /
• 제16권1호
• /
• pp.35-69
• /
• 2014

The impact of competition on pricing has been studied in the context of counterfactual merger analyses where expected optimal prices in a hypothetical monopoly are compared with observed prices in an oligopolistic market. Such analyses would typically assume static decision making by consumers and firms and thus have been applied mostly to data obtained from consumer packed goods such as cereal and soft drinks. However such static modeling approach is not suitable when decision makers are forward looking. When it comes to the markets for durable products with indirect network effects, consumer purchase decisions and firm pricing decisions are inherently dynamic as they take into account future states when making purchase and pricing decisions. Researchers need to take into account the dynamic aspects of decision making both in the consumer side and in the supplier side for such markets. Firms in a two-sided market typically subsidize one side of the market to exploit the indirect network effect. Such pricing behaviors would be more prevalent in competitive markets where firms would try to win over the battle for standard. While such qualitative expectation on the relationship between pricing behaviors and competitive structures could be easily formed, little empirical studies have measured the extent to which the distinct pricing structure in two-sided markets depends on the competitive structure of the market. This paper develops an empirical model to measure the impact of competition on optimal pricing of durable products under indirect network effects. In order to measure the impact of exogenously determined competition among firms on pricing, we compare the equilibrium prices in the observed oligopoly market to those in a hypothetical monopoly market. In computing the equilibrium prices, we account for the forward looking behaviors of consumers and supplier. We first estimate a demand function that accounts for consumers' forward-looking behaviors and indirect network effects. And then, for the supply side, the pricing equation is obtained as an outcome of the Markov Perfect Nash Equilibrium in pricing. In doing so, we utilize numerical dynamic programming techniques. We apply our model to a data set obtained from the U.S. video game console market. The video game console market is considered a prototypical case of two-sided markets in which the platform typically subsidizes one side of market to expand the installed base anticipating larger revenues in the other side of market resulting from the expanded installed base. The data consist of monthly observations of price, hardware unit sales and the number of compatible software titles for Sony PlayStation and Nintendo 64 from September 1996 to August 2002. Sony PlayStation was released to the market a year before Nintendo 64 was launched. We compute the expected equilibrium price path for Nintendo 64 and Playstation for both oligopoly and for monopoly. Our analysis reveals that the price level differs significantly between two competition structures. The merged monopoly is expected to set prices higher by 14.8% for Sony PlayStation and 21.8% for Nintendo 64 on average than the independent firms in an oligopoly would do. And such removal of competition would result in a reduction in consumer value by 43.1%. Higher prices are expected for the hypothetical monopoly because the merged firm does not need to engage in the battle for industry standard. This result is attributed to the distinct property of a two-sided market that competing firms tend to set low prices particularly at the initial period to attract consumers at the introductory stage and to reinforce their own networks and eventually finally to dominate the market.

• 적정기술학회지
• /
• 제7권2호
• /
• pp.127-135
• /
• 2021

전 세계적인 지구온난화의 위협에 대응하고자 세계 각국은 신재생 에너지의 확산, 탄소 배출 감소 등을 추진하고 있다. 또한, UN의 SDGs에도 포함된 기후변화에 맞서기 위한 노력으로 글로벌 자동차 제조사들은 향후 10년내에 전기 자동차로의 전면 전환을 추진하고 있다. 전기자동차는 탄소 배출 감소를 위한 유용한 수단이 될 수 있으나, 충전용 전기를 생산하는 단계에서 발생하는 탄소의 저감을 위해서는 친환경 신재생 에너지를 이용한 발전시스템이 요구된다. 본 연구에서는 아프리카 탄자니아에 설립된 태양광 발전소와 통합된 스마트 전기 모빌리티 운영 시스템에 대한 비전을 제안한다. 아두이노 컴퓨팅 장치를 기반으로 하는 스마트 모니터링 및 통신 기능을 적용하여 전기자동차 또는 전기 오토바이의 배터리 잔존용량, 배터리 상태, 위치, 속도, 고도, 도로 상태 등의 정보를 모니터링한다. 또한, 주변의 독립형 태양광 발전소 인프라와 통신하여 주행가능거리를 예측하고 충전 스케쥴 및 목적지까지의 경로 최적화를 수행하는 시나리오를 제시한다. 제안된 시스템의 구현 가능성은 전기 오토바이의 시험운행을 통해 검증되었다. 탄자니아에서 운영될 전기 모빌리티 시스템은 현지의 환경과 특성을 고려하여 친환경성, 경제성, 운용 용이성, 호환성 등의 요소가 고려되어야 한다. 본 연구에서 제안하는 스마트 전기 모빌리티 운영 시스템은 SDGs의 이행을 위한 중요한 기반이 될 수 있을 것이다.