• The KIPS Transactions:PartB
• /
• v.10B no.6
• /
• pp.665-672
• /
• 2003

This paper describes a speech synthesis technique by concatenating unit phoneme. At that time, a major problem is that discontinuity is happened from connection part between unit phonemes, especially from connection part between unit phonemes recorded by different persons. To solve the problem, this paper uses clustered diphone, and proposes a spectral smoothing technique, not only using formant trajectory and distribution characteristic of spectrum but also reflecting human's acoustic characteristic. That is, the proposed technique performs unit phoneme clustering using distribution characteristic of spectrum at connection part between unit phonemes and decides a quantity and a scope for the smoothing by considering human's acoustic characteristic at the connection part of unit phonemes, and then performs the spectral smoothing using weights calculated along a time axes at the border of two diphones. The proposed technique removes the discontinuity and minimizes the distortion which can be occurred by spectrum smoothing. For the purpose of the performance evaluation, we test on five hundred diphones which are extracted from twenty sentences recorded by five persons, and show the experimental results.

• Journal of Korea Multimedia Society
• /
• v.11 no.9
• /
• pp.1302-1309
• /
• 2008

Traditional clustering methods, like k-means or fuzzy clustering, are prototype-based methods which are applicable only to convex clusters. On the other hand, spectral clustering tries to find clusters only using local similarity information. Its ability to handle concave clusters has gained the popularity recent years together with support vector machine (SVM) which is a kernel-based classification method. However, as is in SVM, the kernel width plays an important role and has a great impact on the result. Several methods are proposed to decide it automatically, it is still determined based on heuristics. In this paper, we proposed an adaptive method deciding the kernel width based on distance histogram. The proposed method is motivated by the fact that the affinity matrix should be formed into a block diagonal matrix to generate the best result. We use the tradition Euclidean distance together with the random walk distance, which make it possible to form a more apparent block diagonal affinity matrix. Experimental results show that the proposed method generates more clear block structured affinity matrix than the existing one does.

• The Korean Journal of Applied Statistics
• /
• v.33 no.4
• /
• pp.511-526
• /
• 2020

Single-cell RNA-sequencing (scRNA-seq) data consists of each cell's RNA expression extracted from large populations of cells. One main purpose of using scRNA-seq data is to identify inter-cellular heterogeneity. However, scRNA-seq data pose statistical challenges when applying traditional clustering methods because they have many missing values and high level of noise due to technical and sampling issues. In this paper, motivated by analyzing scRNA-seq data, we propose a novel spectral-based clustering method by imposing different weights on genes when computing a similarity between cells. Assigning weights on genes and clustering cells are performed simultaneously in the proposed clustering framework. We solve the proposed non-convex optimization using an iterative algorithm. Both real data application and simulation study suggest that the proposed clustering method better identifies underlying clusters compared with existing clustering methods.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2018.05a
• /
• pp.437-438
• /
• 2018

문서 군집화 성능을 높이기 위한 한 방법으로 차원 축소를 적용한 문서 벡터로 군집화를 실시하는 방법이 있다. 본 발표에서는 특이값 분해(SVD), 커널 주성분 분석(Kernel PCA), Doc2Vec 등의 차원 축소 기법을, K-평균 군집화(K-means clustering), 계층적 병합 군집화(hierarchical agglomerative clustering), 스펙트럼 군집화(spectral clustering)에 적용하고, 그 성능을 비교해 본다.

• The Korean Journal of Applied Statistics
• /
• v.21 no.3
• /
• pp.547-560
• /
• 2008

It is well known that spectral density function determines auto-covariance function of stationary time-series data and smoothed periodogram is a consistent estimator of spectral density function. Recently, Kim and Park (2007) showed that smoothed- periodogram based distances performs very well for the classification. In this paper, we introduce classification methods with smoothed periodogram and apply the approaches to the monthly precipitation measurements obtained from January, 1987 through December, 2007 at 22 locations in South Korea.

• Journal of the Korean Data and Information Science Society
• /
• v.27 no.3
• /
• pp.621-627
• /
• 2016

A certain professional baseball team tends to be very weak against another particular team. For example, S team, the strongest team in Korea, is relatively weak to H team. In this paper, we carried out clustering the Korean baseball teams based on the records against the team S to investigate whether the pattern of the record of the team H is different from those of the other teams. The technique we have employed is 'time series clustering', or more specifically 'categorical time series clustering'. Three methods have been considered in this paper: (i) distance based method, (ii) genetic sequencing method and (iii) periodogram method. Each method has its own advantages and disadvantages to handle categorical time series, so that it is recommended to draw conclusion by considering the results from the above three methods altogether in a comprehensive manner.

• The Journal of the Acoustical Society of Korea
• /
• v.35 no.4
• /
• pp.303-309
• /
• 2016

This paper proposes a novel approach to the problem of audio signal clustering using a stacked autoencoder. The proposed stacked autoencoder learns an efficient representation for the input signal, enables clustering constituent signals with similar characteristics, and therefore the original sources can be separated based on the clustering results. STFT (Short-Time Fourier Transform) is performed to extract time-frequency spectrum, and rectangular windows at all the possible locations are used as input values to the autoencoder. The outputs at the middle, encoding layer, are used to cluster the rectangular windows and the original sources are separated by the Wiener filters derived from the clustering results. Source separation experiments were carried out in comparison to the conventional NMF (Non-negative Matrix Factorization), and the estimated sources by the proposed method well represent the characteristics of the orignal sources as shown in the time-frequency representation.

• Proceedings of the Korean Information Science Society Conference
• /
• 2005.07a
• /
• pp.892-894
• /
• 2005

Graph partitioning provides an important tool for data clustering, but is an NP-hard combinatorial optimization problem. Spectral clustering where the clustering is performed by the eigen-decomposition of an affinity matrix [1,2]. This is a popular way of solving the graph partitioning problem. On the other hand, semidefinite relaxation, is an alternative way of relaxing combinatorial optimization. issuing to a convex optimization[4]. In this paper we present a semidefinite programming (SDP) approach to graph equi-partitioning for clustering and then we use eigen-decomposition to obtain an optimal partition set. Therefore, the method is referred to as semidefinite spectral clustering (SSC). Numerical experiments with several artificial and real data sets, demonstrate the useful behavior of our SSC. compared to existing spectral clustering methods.

• Phonetics and Speech Sciences
• /
• v.11 no.3
• /
• pp.39-47
• /
• 2019

We propose a method to improve the performance of spontaneous speech recognizers by extending their phone set using speech data. In the proposed method, we first extract variable-length phoneme-level segments from broadcast speech signals, and convert them to fixed-length latent vectors using an long short-term memory (LSTM) classifier. We then cluster acoustically similar latent vectors and build a new phone set by choosing the number of clusters with the lowest Davies-Bouldin index. We also update the lexicon of the speech recognizer by choosing the pronunciation sequence of each word with the highest conditional probability. In order to analyze the acoustic characteristics of the new phone set, we visualize its spectral patterns and segment duration. Through speech recognition experiments using a larger training data set than our own previous work, we confirm that the new phone set yields better performance than the conventional phoneme-based and grapheme-based units in both spontaneous speech recognition and read speech recognition.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.40 no.11
• /
• pp.963-971
• /
• 2012

In this paper, development of optimal design tools for wing structure is described including multi load spectra condition and fatigue analysis. Two dimensional CFD result are used for calculating aerodynamic force. Design variables are composed of a number of rib and spar, positions, and thickness of each structural member. The mission profile for fatigue analysis is composed based upon the results of CFD analysis, the flight-by-flight spectra method, the excessive curves for gust loads. Minor's rule was used to deal with multi-load condition. Stress analysis and fatigue analysis are performed to calculate objective functions. Particle Swarm Optimization(PSO) algorithm was used to apply to problems which have dozens of design variables.