• 한국통신학회논문지
• /
• 제27권6A호
• /
• pp.517-523
• /
• 2002

이 논문에서는 국소 최적 순위 검파기의 점수 함수의 합과 가중합의 완전한 골을 유도하였다. 순위통계량과 부호통계량에 바탕을 둔 검파기의 점근 성능 특성은 점수 함수의 합과 가 중합으로부터 얻어지기 때문에, 합과 가중합은 매우 중요하나, 이들을 구하기 위해서는 상당한 수학적 조작이 필요하다. 이 논문에서 다루어진 점수 함수는 순위통계량에 바탕을 둔 것들 뿐 아니라 절대값 순위통계량과 부호통계량에 바탕을 둔 것들을 포함한다. 따라서, 이 논문의 결과를 써서 여러 가지 잡음 모형에서 쓸모 있는 검 파기들의 점근 성능을 쉽게 구할 수 있다.

• ETRI Journal
• /
• 제41권2호
• /
• pp.235-241
• /
• 2019

In this paper, we propose a rank-weighted reconstruction feature to improve the robustness of a feed-forward deep neural network (FFDNN)-based acoustic model. In the FFDNN-based acoustic model, an input feature is constructed by vectorizing a submatrix that is created by slicing the feature vectors of frames within a context window. In this type of feature construction, the appropriate context window size is important because it determines the amount of trivial or discriminative information, such as redundancy, or temporal context of the input features. However, we ascertained whether a single parameter is sufficiently able to control the quantity of information. Therefore, we investigated the input feature construction from the perspectives of rank and nullity, and proposed a rank-weighted reconstruction feature herein, that allows for the retention of speech information components and the reduction in trivial components. The proposed method was evaluated in the TIMIT phone recognition and Wall Street Journal (WSJ) domains. The proposed method reduced the phone error rate of the TIMIT domain from 18.4% to 18.0%, and the word error rate of the WSJ domain from 4.70% to 4.43%.

• 응용통계연구
• /
• 제7권2호
• /
• pp.79-88
• /
• 1994

가중 로그 순위 검정은 생존분석에서 두 집단의 차이를 검정하는데 가장 많이 사용되는 검정이다. 생존실험에서 실험 초기에는 별다른 차이를 보이지 않다가 실험 중반을 넘어가며 집단간의 차이를 보이는 경우가 많은데 이러한 경우에 기존에 제시되어 있는 여러 가중 로그 순위 검정은 비효율적이다. 이 문제에 적당한 새로운 가중 로그 순위검정을 제시하며 예를 통해서 제시된 검정의 유용성을 살펴보고 모의실험을 통해 이들의 검정력도 조사한다. 제시된 검정은 후기 차이 검출에 매우 효율적이다.

• 의료ㆍ복지 건축 : 한국의료복지건축학회 논문집
• /
• 제27권2호
• /
• pp.25-37
• /
• 2021

Purpose: The main purpose of this study is to develop an evaluation system using the weighted-values of various users and experts for the public space to apply Universal Design, and additionally to find out the commonalities and differences by comparing the importance of evaluation indicators between users and expert groups. Method: A one-sample t-test was conducted to verify that the components of the public space to universal design application are suitable as evaluation indicators, and AHP(analytic hierarchy process) was performed to derive weight-values for the evaluation system. Results: The importance-values for the total 23 facilities to be used as evaluation indicators were derived by multiplying the weighted-values of each sector, domain, and facility by the disabled, non-disabled, and experts. To summarize the results of overall importance-values derived from the AHP, The disabled showed high-rank weighted-values in facilities of building sector > park & recreation sector > cross domain and low-rank weighted-values for sidewalk and roadway domain. The non-disabled showed high-rank weighted-values in facilities of park & recreation sector > roadway domain > building sector > cross domain and low-rank weighted-values for sidewalk domain. Experts mainly showed high-rank weighted-values in the cross domain and in facilities related to entry and movement to the target space in all sectors and domains. However, it showed moderate importance-values in the sanitary space. The disabled who are restricted to movement have a high demand for universal design in buildings consisting of vertical moving line, and non-disabled people who are not limited to physical movement have a high demand for universal design in parks and recreation sector for increased leisure time. It means that experts are important to recognize the principles of making space because they value cross domain and the key spaces and facilities for suitable the purpose of use. In addition, it can be inferred that non-disabled people have a higher demand for safety than disabled people due to their high importance in roadway domain and facilities of safety and disaster prevention. Implications: The significance of this study is the establishment of a quantitative universal design evaluation system for public spaces considering the different perspectives of the disabled and the non-disabled.

• 대한수학회지
• /
• 제56권6호
• /
• pp.1503-1514
• /
• 2019

A rank 1 matrix has a factorization as $uv^t$ for vectors u and v of some orders. The arctic rank of a rank 1 matrix is the half number of nonzero entries in u and v. A matrix of rank k can be expressed as the sum of k rank 1 matrices, a rank 1 decomposition. The arctic rank of a matrix A of rank k is the minimum of the sums of arctic ranks of the rank 1 matrices over all rank 1 decomposition of A. In this paper we obtain characterizations of the linear operators that strongly preserve the symmetric arctic ranks of symmetric matrices over nonnegative reals.

• Current Optics and Photonics
• /
• 제5권4호
• /
• pp.431-443
• /
• 2021

Sparse unmixing has been proven to be an effective method for hyperspectral unmixing. Hyperspectral images contain rich spectral and spatial information. The means to make full use of spectral information, spatial information, and enhanced sparsity constraints are the main research directions to improve the accuracy of sparse unmixing. However, many algorithms only focus on one or two of these factors, because it is difficult to construct an unmixing model that considers all three factors. To address this issue, a novel algorithm called multiview-based spectral weighted and low-rank row-sparsity unmixing is proposed. A multiview data set is generated through spectral partitioning, and then spectral weighting is imposed on it to exploit the abundant spectral information. The row-sparsity approach, which controls the sparsity by the l_2,0 norm, outperforms the single-sparsity approach in many scenarios. Many algorithms use convex relaxation methods to solve the l_2,0 norm to avoid the NP-hard problem, but this will reduce sparsity and unmixing accuracy. In this paper, a row-hard-threshold function is introduced to solve the l_2,0 norm directly, which guarantees the sparsity of the results. The high spatial correlation of hyperspectral images is associated with low column rank; therefore, the low-rank constraint is adopted to utilize spatial information. Experiments with simulated and real data prove that the proposed algorithm can obtain better unmixing results.

• ETRI Journal
• /
• 제35권4호
• /
• pp.687-696
• /
• 2013

The development of proxy caching is essential in the area of video-on-demand (VoD) to meet users' expectations. VoD requires high bandwidth and creates high traffic due to the nature of media. Many researchers have developed proxy caching models to reduce bandwidth consumption and traffic. Proxy caching keeps part of a media object to meet the viewing expectations of users without delay and provides interactive playback. If the caching is done continuously, the entire cache space will be exhausted at one stage. Hence, the proxy server must apply cache replacement policies to replace existing objects and allocate the cache space for the incoming objects. Researchers have developed many cache replacement policies by considering several parameters, such as recency, access frequency, cost of retrieval, and size of the object. In this paper, the Weighted-Rank Cache replacement Policy (WRCP) is proposed. This policy uses such parameters as access frequency, aging, and mean access gap ratio and such functions as size and cost of retrieval. The WRCP applies our previously developed proxy caching model, Hot-Point Proxy, at four levels of replacement, depending on the cache requirement. Simulation results show that the WRCP outperforms our earlier model, the Dual Cache Replacement Policy.

• 대한임베디드공학회논문지
• /
• 제10권2호
• /
• pp.65-71
• /
• 2015

It has been known that GMM (Gaussian Mixture Model) based speaker identification systems using ML (Maximum Likelihood) and WMR (Weighting Model Rank) demonstrate very high performances. However, such systems are not so effective under practical environments, in terms of real time processing, because of their high calculation costs. In this paper, we propose a new speaker-pruning algorithm that effectively reduces the calculation cost. In this algorithm, we select 20% of speaker models having higher likelihood with a part of input speech and apply MWMR (Modified Weighted Model Rank) to these selected speaker models to find out identified speaker. To verify the effectiveness of the proposed algorithm, we performed speaker identification experiments using TIMIT database. The proposed method shows more than 60% improvement of reduced processing time than the conventional GMM based system with no pruning, while maintaining the recognition accuracy.

• Korean Journal of Mathematics
• /
• 제28권4호
• /
• pp.791-802
• /
• 2020

We obtain a two variable generating function for the number of triangular partitions. Using this generating function, we study arithmetic properties of a certain weighted count of triangular partitions. Finally, we introduce a rank-type function for triangular partitions, which gives a combinatorial explanation for a triangular partition congruence.

• Kyungpook Mathematical Journal
• /
• 제43권4호
• /
• pp.605-616
• /
• 2003

In this paper, some rank equalities related to generalized inverses $A^{(2)}_{T,S}$ of a matrix are presented. As applications, a variety of rank equalities related to the M-P inverse, the Drazin inverse, the group inverse, the weighted M-P inverse, the Bott-Duffin inverse and the generalized Bott-Duffin inverse are established.