Researcher Profile and Settings

Master

Affiliation (Master)

• Faculty of Information Science and Technology　Computer Science and Information Technology Mathematical Science

Affiliation (Master)

• Faculty of Information Science and Technology　Computer Science and Information Technology Mathematical Science

researchmap

Profile and Settings

Degree

• (BLANK)（Hokkaido University）
• (BLANK)（Hokkaido University）

Profile and Settings

• Name (Japanese)

  Imai

• Name (Kana)

  Hideyuki

• Name

  200901009464751373

Achievement

Research Interests

• 統計学   Statistics   

Research Areas

• Informatics / Statistical science

Published Papers

• Kernelized Supervised Laplacian Eigenmap for Visualization and Classification of Multi-Label Data

  Mariko Tai, Mineichi Kudo, Akira Tanaka, Hideyuki Imai, Keigo Kimura

  Pattern Recognition 123 108399 - 108399 0031-3203 2022/03
• Kernel-Based Regressors Equivalent to Stochastic Affine Estimators

  Akira TANAKA, Masanari NAKAMURA, Hideyuki IMAI

  IEICE Transactions on Information and Systems E105.D (1) 116 - 122 0916-8532 2022/01/01
• Comparison of dimension reduction methods on fatty acids food source study.

  Yifan Chen, Yusuke Miura, Toshihiro Sakurai, Zhen Chen, Rojeet Shrestha, Sota Kato, Emiko Okada, Shigekazu Ukawa, Takafumi Nakagawa, Koshi Nakamura, Akiko Tamakoshi, Hitoshi Chiba, Hideyuki Imai, Hiroyuki Minami, Masahiro Mizuta, Shu-Ping Hui

  Scientific reports 11 (1) 18748 - 18748 2021/09/21 

  Serum fatty acids (FAs) exist in the four lipid fractions of triglycerides (TGs), phospholipids (PLs), cholesteryl esters (CEs) and free fatty acids (FFAs). Total fatty acids (TFAs) indicate the sum of FAs in them. In this study, four statistical analysis methods, which are independent component analysis (ICA), factor analysis, common principal component analysis (CPCA) and principal component analysis (PCA), were conducted to uncover food sources of FAs among the four lipid fractions (CE, FFA, and TG + PL). Among the methods, ICA provided the most suggestive results. To distinguish the animal fat intake from endogenous fatty acids, FFA variables in ICA and factor analysis were studied. ICA provided more distinct suggestions of FA food sources (endogenous, plant oil intake, animal fat intake, and fish oil intake) than factor analysis. Moreover, ICA was discovered as a new approach to distinguish animal FAs from endogenous FAs, which will have an impact on epidemiological studies. In addition, the correlation coefficients between a published dataset of food FA compositions and the loading values obtained in the present ICA study suggested specific foods as serum FA sources. In conclusion, we found that ICA is a useful tool to uncover food sources of serum FAs.
• Principal Component Clustered Factors for Determining Study Performance in Computer Programming Class

  Unhawa Ninrutsirikun, Hideyuki Imai, Bunthit Watanapa, Chonlameth Arpnikanondt

  Wireless Personal Communications 115 (4) 2897 - 2916 0929-6212 2020/12 [Refereed]
  
• Shift Invariance Property of a Non-Negative Matrix Factorization

  Hideyuki IMAI

  IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences E103.A (2) 580 - 581 0916-8508 2020/02/01 [Refereed]
  
• A Fast Cross-Validation Algorithm for Kernel Ridge Regression by Eigenvalue Decomposition

  TANAKA Akira, IMAI Hideyuki

  IEICE Transactions on Electronics, Communications and Computer Sciences E102-A (9) 1317 - 1320 2019/09 [Refereed][Not invited]
  
• Determination of Serum 25-Hydroxyvitamin D₃ by LC/MS/MS and Its Monthly Variation in Sapporo Indoor Workers

  Hiroaki OKABE, Chikara SHIMIZU, Maiko YAMAMOTO, Rei KIKUCHI, Akiko MINAMI, Yi-Fan CHEN, Hideyuki IMAI, Masahiro MIZUTA, Zhen CHEN, Hitoshi CHIBA, Shu-Ping HUI

  Analytical Sciences 34 (9) 1043 - 1047 0910-6340 2018/09/10 [Invited]
   

  ビタミンDの代謝物である25-ヒドロキシビタミンD3(25(OH)D3)の血清中濃度に対し、年齢や性別が及ぼす影響を明らかにするため、液体クロマトグラフィータンデム質量分析法(LC/MS/MS)による25(OH)D3の測定法を検討した。屋内労働者の男性11名(平均34.9±10.4歳)と女性11名(平均34.2±6.5歳)を対象に、25(OH)D3の血清中濃度を1年間にわたり毎月測定した。その結果、25(OH)D3の血清中濃度の全平均は12.9±4.7ng/mLであった。女性の25(OH)D3濃度は男性より有意に低かった(4.0±5.0 vs.11.9±4.3ng/mL)。また、男女共、血清中の25(OH)D3濃度は、2ヵ月の遅れを伴って紫外線B波照射(または日照時間)と強く相関することが判明した(それぞれr2=0.8477、0.7384)。さらに、男性の25(OH)D3濃度の月間変動は女性より著しいことも明らかになった。
• An Online Self-Constructive Normalized Gaussian Network with Localized Forgetting

  Jana Backhus, Ichigaku Takigawa, Hideyuki Imai, Mineichi Kudo, Masanori Sugimoto

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E100A (3) 865 - 876 1745-1337 2017/03 [Refereed][Not invited]
   

  In this paper, we introduce a self-constructive Normalized Gaussian Network (NGnet) for online learning tasks. In online tasks, data samples are received sequentially, and domain knowledge is often limited. Then, we need to employ learning methods to the NGnet that possess robust performance and dynamically select an accurate model size. We revise a previously proposed localized forgetting approach for the NGnet and adapt some unit manipulation mechanisms to it for dynamic model selection. The mechanisms are improved for more robustness in negative interference prone environments, and a new merge manipulation is considered to deal with model redundancies. The effectiveness of the proposed method is compared with the previous localized forgetting approach and an established learning method for the NGnet. Several experiments are conducted for a function approximation and chaotic time series forecasting task. The proposed approach possesses robust and favorable performance in different learning situations over all testbeds.
• Theoretical Analyses on 2-Norm-Based Multiple Kernel Regressors

  TANAKA Akira, IMAI Hideyuki

  IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences 電子情報通信学会(The Institute of Electronics, Information and Communication Engineers / IEICE) 100 (3) 877 - 887 1745-1337 2017 

  The solution of the standard 2-norm-based multiple kernel regression problem and the theoretical limit of the considered model space are discussed in this paper. We prove that 1) The solution of the 2-norm-based multiple kernel regressor constructed by a given training data set does not generally attain the theoretical limit of the considered model space in terms of the generalization errors, even if the training data set is noise-free, 2) The solution of the 2-norm-based multiple kernel regressor is identical to the solution of the single kernel regressor under a noise free setting, in which the adopted single kernel is the sum of the same kernels used in the multiple kernel regressor; and it is also true for a noisy setting with the 2-norm-based regularizer. The first result motivates us to develop a novel framework for the multiple kernel regression problems which yields a better solution close to the theoretical limit, and the second result implies that it is enough to use the single kernel regressors with the sum of given multiple kernels instead of the multiple kernel regressors as long as the 2-norm based criterion is used.

• Whisper to normal speech conversion using pitch estimated from spectrum.

  Hideaki Konno, Mineichi Kudo, Hideyuki Imai, Masanori Sugimoto

  Speech Commun. 83 10 - 20 2016
• Dense core model for cohesive subgraph discovery

  Sadamori Koujaku, Ichigaku Takigawa, Mineichi Kudo, Hideyuki Imai

  SOCIAL NETWORKS 44 143 - 152 0378-8733 2016/01 [Refereed][Not invited]
   

  Discovery of cohesive subgraphs is an important issue in social network analysis. As representative cohesive subgraphs, pseudo cliques have been developed by relaxing the perfection of cliques. By enumerating pseudo clique subgraphs, we can find some structures of interest such as a star-like structure. However, a little more complicated structures such as a core/periphery structure is still hard to be found by them. Therefore, we propose a novel pseudo clique called p-dense core and show the connection with the other pseudo cliques. Moreover, we show that a set of p-dense core subgraphs gives an optimal solution in a graph partitioning problem. Several experiments on real-life networks demonstrated the effectiveness for cohesive subgraph discovery. (C) 2015 Elsevier B.V. All rights reserved.
• Online EM for the Normalized Gaussian Network with Weight-Time-Dependent Updates

  Jana Backhus, Ichigaku Takigawa, Hideyuki Imai, Mineichi Kudo, Masanori Sugimoto

  NEURAL INFORMATION PROCESSING, ICONIP 2016, PT IV 9950 538 - 546 0302-9743 2016 [Refereed][Not invited]
   

  In this paper, we propose a weight-time-dependent (WTD) update approach for an online EM algorithm applied to the Normalized Gaussian network (NGnet). WTD aims to improve a recently proposed weight-dependent (WD) update approach by Celaya and Agostini. First, we discuss the derivation of WD from an older time-dependent (TD) update approach. Then, we consider additional aspects to improve WD, and by including them we derive the new WTD approach from TD. The difference between WD and WTD is discussed, and some experiments are conducted to demonstrate the effectiveness of the proposed approach. WTD succeeds in improving the learning performance for a function approximation task with balanced and dynamic data distributions.
• Reducing Redundancy with Unit Merging for Self-constructive Normalized Gaussian Networks

  Jana Backhus, Ichigaku Takigawa, Hideyuki Imai, Mineichi Kudo, Masanori Sugimoto

  ARTIFICIAL NEURAL NETWORKS AND MACHINE LEARNING - ICANN 2016, PT I 9886 444 - 452 0302-9743 2016 [Refereed][Not invited]
   

  In this paper, a Normalized Gaussian Network (NGnet) is introduced for online sequential learning that uses unit manipulation mechanisms to build the network model self-constructively. Several unit manipulation mechanisms have been proposed for online learning of an NGnet. However, unit redundancy still exists in the network model. We propose a merge mechanism for such redundant units, and change its overlap calculation in order to improve the identification accuracy of redundant units. The effectiveness of the proposed approach is demonstrated in a function approximation task with balanced and imbalanced data distributions. It succeeded in reducing the model complexity around 11% on average while keeping or even improving learning performance.
• Ensemble and Multiple Kernel Regressors: Which Is Better?

  TANAKA Akira, TAKEBAYASHI Hirofumi, TAKIGAWA Ichigaku, IMAI Hideyuki, KUDO Mineichi

  IEICE Transactions on Fundamentals of Electronics, Communications and Computer Science E98-A (11) 2315 - 2423 2015/11 [Refereed][Not invited]
  
• Parametric Wiener Filter with Linear Constraints for Unknown Target Signals

  Akira Tanaka, Hideyuki Imai

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E97A (1) 322 - 330 0916-8508 2014/01 [Refereed][Not invited]
   

  In signal restoration problems, we expect to improve the restoration performance with a priori information about unknown target signals. In this paper, the parametric Wiener filter with linear constraints for unknown target signals is discussed. Since the parametric Wiener filter is usually defined as the minimizer of the criterion not for the unknown target signal but for the,filter, it is difficult to impose constraints for the unknown target signal in the criterion. To overcome this difficulty, we introduce a criterion for the parametric Wiener filter defined for the unknown target signal whose minimizer is equivalent to the solution obtained by the original formulation. On the basis of the newly obtained criterion, we derive a closed-form solution for the parametric Wiener filter with linear constraints.
• Extended Analyses for an Optimal Kernel in a Class of Kernels with an Invariant Metric

  Akira Tanaka, Ichigaku Takigawa, Hideyuki Imai, Mineichi Kudo

  STRUCTURAL, SYNTACTIC, AND STATISTICAL PATTERN RECOGNITION 7626 345 - 353 0302-9743 2012/11 [Refereed][Not invited]
   

  Learning based on kernel machines is widely known as a powerful tool for various fields of information science such as pattern recognition and regression estimation. An appropriate model selection is required in order to obtain desirable learning results. In our previous work, we discussed a class of kernels forming a nested class of reproducing kernel Hilbert spaces with an invariant metric and proved that the kernel corresponding to the smallest reproducing kernel Hilbert space, including an unknown true function, gives the best model. In this paper, we relax the invariant metric condition and show that a similar result is obtained when a subspace with an invariant metric exists.
• Kernel-Induced Sampling Theorem

  Tanaka Akira, Imai Hideyuki, Miyakoshi Masaaki

  IEEE Transactions on Signal Processing IEEE - Institute of Electrical and Electronics Engineers 58 (7) 3569 - 3577 1053-587X 2010/07 [Refereed][Not invited]
   

  A perfect reconstruction of functions in a reproducing kernel Hilbert space from a given set of sampling points is discussed. A necessary and sufficient condition for the corresponding reproducing kernel and the given set of sampling points to perfectly recover the functions is obtained in this paper. The key idea of our work is adopting the reproducing kernel Hilbert space corresponding to the Gramian matrix of the kernel and the given set of sampling points as the range space of a sampling operator and considering the orthogonal projector, defined via the range space, onto the closed linear subspace spanned by the kernel functions corresponding to the given sampling points. We also give an error analysis of a reconstructed function by incomplete sampling points.
• Probably correct k-nearest neighbor search in high dimensions

  Jun Toyama, Mineichi Kudo, Hideyuki Imai

  PATTERN RECOGNITION 43 (4) 1361 - 1372 0031-3203 2010/04 [Refereed][Not invited]
   

  A novel approach for k-nearest neighbor (k-NN) searching with Euclidean metric is described. It is well known that many sophisticated algorithms cannot beat the brute-force algorithm when the dimensionality is high. In this study, a probably correct approach, in which the correct set of k-nearest neighbors is obtained in high probability, is proposed for greatly reducing the searching time. We exploit the marginal distribution of the k th nearest neighbors in low dimensions, which is estimated from the stored data (an empirical percentile approach). We analyze the basic nature of the marginal distribution and show the advantage of the implemented algorithm, which is a probabilistic variant of the partial distance searching. Its query time is sublinear in data size n, that is, O(mn delta) with S=o(1) in n and delta <= 1, for any fixed dimension m. (C) 2009 Elsevier Ltd. All rights reserved.
• Integrated kernels and their properties

  Akira Tanaka, Hideyuki Imai, Mineichi Kudo, Masaaki Miyakoshi

  PATTERN RECOGNITION 40 (11) 2930 - 2938 0031-3203 2007/11 [Not refereed][Not invited]
   

  Kernel machines are widely considered to be powerful tools in various fields of information science. By using a kernel, an unknown target is represented by a function that belongs to a reproducing kernel Hilbert space (RKHS) corresponding to the kernel. The application area is widened by enlarging the RKHS such that it includes a wide class of functions. In this study, we demonstrate a method to perform this by using parameter integration of a parameterized kernel. Some numerical experiments show that the unresolved problem of finding a good parameter can be neglected. (c) 2007 Pattern Recognition Society. Published by Elsevier Ltd. All rights reserved.
• A Unified Framework of Subspace Identification for D.O.A. Estimation

  TANAKA Akira, IMAI Hideyuki, MIYAKOSHI Masaaki

  IEICE Trans. Fundamentals, A 一般社団法人電子情報通信学会 E90-A (2) 419 - 428 0916-8508 2007/02 [Refereed][Not invited]
   

  In D.O.A. estimation, identification of the signal and the noise subspaces plays an essential role. This identification process was traditionally achieved by the eigenvalue decomposition (EVD) of the spatial correlation matrix of observations or the generalized eigenvalue decomposition (GEVD) of the spatial correlation matrix of observations with respect to that of an observation noise. The framework based on the GEVD is not always an extension of that based on the EVD, since the GEVD is not applicable to the noise-free case which can be resolved by the framework based on the EVD. Moreover, they are not applicable to the case in which the spatial correlation matrix of the noise is singular. Recently, a quotient-singular-value-decomposition-based framework, that can be applied to problems with singular noise correlation matrices, is introduced for noise reduction. However, this framework also can not treat the noise-free case. Thus, we do not have a unified framework of the identification of these subspaces. In this paper, we show that a unified framework of the identification of these subspaces is realized by the concept of proper and improper eigenspaces of the spatial correlation matrix of the noise with respect to that of observations.
• Image Restoration in Singular Observation Processes Based on Complementary Space Component Estimation

  A. Tanaka, H. Imai, M. Miyakoshi

  Electronics and Communications in Japan, Part 2 88 (8) 54 - 65 2005 [Refereed][Not invited]
  
• Projection learning based kernel machine design using series of monotone increasing reproducing kernel Hilbert spaces

  A Tanaka, Takigawa, I, H Imai, M Kudo, M Miyakoshi

  KNOWLEDGE-BASED INTELLIGENT INFORMATION AND ENGINEERING SYSTEMS, PT 1, PROCEEDINGS 3213 1058 - 1064 0302-9743 2004 [Refereed][Not invited]
   

  Kernel machines are widely known as powerful tools for various fields of information science. In general, they are designed based on a generalization criterion related to the complexity of the model and intuitive but ad hoe philosophy such as maximal margin principle shown in SVM. On the other hand, the project ion learning scheme was proposed in the field of neural networks. In the projection learning, the generalization ability is evaluated by the distance between the unknown target function and the estimated one. In,this paper, we construct projection learning based kernel machines and propose a method of making a kernel function that has necessary representability for the task. The method is reduced to a selection of an appropriate reproducing kernel Hilbert space from a series of monotone increasing subspaces. We also verify the efficacy of the proposed method by numerical examples.
• Global optimization of a perturbed problem using interval analysis)

  Waweru Mwangi, Hideyuki Imai, Hirohito Sakurai, Yoshiharu Sato

  Bulletin of the Computational Statistics of Japan 日本計算機統計学会 16 (2) 167 - 167 2004
• Improvements of goodness-of-fit statistics for sparse multinomials based on normalizing transformations

  Nobuhiro Taneichi, Yuri Sekiya, Hideyuki Imai

  Annals of the Institute of Statistical Mathematics 55 (4) 831 - 848 0020-3157 2003/12 [Refereed][Not invited]
  
• Image Restoration in Singular Observation Processes Based on Complementary Space Component Estimation

  TANAKA Akira, IMAI Hideyuki, MIYAKOSHI Masaaki

  The Transactions of the Institute of Electronics,Information and Communication Engineers. 一般社団法人電子情報通信学会 Vol.J86-D-II (12) 1745 - 1755 0915-1923 2003/12 [Refereed][Not invited]
   

  画像復元問題とは,何らかの観測過程により劣化を受けた画像から未知の原画像を推定する問題であり,これまで様々な復元法が提案されている.従来手法の多くでは,観測過程は正則であるか,若しくはそれに近い状況であることが前提となっているが,一般に観測過程が正則である保証はなく,そのような観測過程に対する復元性能は十分とはいえない.特異な観測過程に対する復元性能の低下は,観測過程によって消失した零空間に関連する成分の推定が困難であることに起因する.ところで,近年,画像信号の性質や表現に関する研究も数多くなされており,その中で得られた知見の一つとして,画像の差信号がおおむねラプラス分布に従うということが報告されている.本論文では,この画像の確率的性質に基づき,特異な観測過程によって消失した画像成分の推定を積極的に行うことにより,高精度の復元を行う手法を提案することを目的とする.また,数値実験により提案手法の有効性を検証する.
• Multidimensional global optimization using interval slopes

  RW Mwangi, H Imai, Y Sato

  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES E86A (11) 2836 - 2843 0916-8508 2003/11 [Refereed][Not invited]
   

  The knowledge of a good enclosure of the range of a function over small interval regions allows us to avoid convergence of optimization algorithms to a non-global point(s). We used interval slopes f [X, x] to check for monotonicity and integrated their derivative forms g[X, x], x is an element of X by quadratic and Newton methods to obtain narrow enclosures. In order to include boundary points in the search for the optimum point(s), we expanded the initial box by a small width on each dimension. These procedures resulted in an improvement in the algorithm proposed by Hansen [4].
• TESTS OF DIMENSIONALITY FOR MANOVA UNDER MULTIVARIATE t POPULATIONS

  Yoshida Kiyotaka, Imai Hideyuki, Sato Yoshiharu

  Bulletin of the Computational Statistics of Japan 日本計算機統計学会 16 (1) 21 - 29 0914-8930 2003 

  For tests of dimensionality in multivariate analysis of variance (MANOVA), three types of test criteria (Likelihood-Ratio-type, Lawley-Hotelling-type and Bartlett-Nanda-Pillai-type) are popular. However, as is well known, these criteria depend on nuisance parameters. Yoshida et al. (2002) suggested new criteria, which are upper bound for null distribution of the above-mentionted three types criteria and do not depend on the nuisance parameters under elliptical populations. Furthermore, it was proved that the upper bounds are upper limits for null distributions of LR-type and LH-type. In this paper, we consider tests of dimensionality under multivariate t populations, and show that the upper bounds are also upper limits for null distributions of LR-type and LH-type.
• Image Restoration by the Family of Signal Adaptive Parametric Projection Filters

  A. Tanaka, H. Imai, M. Miyakoshi

  Electronics and Communications in Japan, Part 3 86 (3) 77 - 86 2003 [Refereed][Not invited]
  
• On Duality of Constrained Least-Squares Restoration and Parametric Wiener Restoration

  田中章, 今井英幸, 宮腰政明

  第17回DSPシンポジウム講演論文集 A2-1  2002/11 [Not refereed][Not invited]
  
• On Duality of Formulations in Image Restoration Problems

  A. Tanaka, H. Imai, M. Miyakoshi

  Proceedings of the 4th IASTED International Conference Signal and Image Processing 242 - 247 2002/08 [Refereed][Not invited]
  
• Relations and Differences Between Classical and Modern Formulations in Image Restoration Problems

  A. Tanaka, H. Imai, M. Miyakoshi

  Proceedings of the 6th World Multiconference on Systemics, Cybernetics and Informatics XVI 324 - 328 2002/07 [Refereed][Not invited]
  
• Separable Kernel Expression of Regularized MP-Inverse for Separable Kernel Operators

  TANAKA Akira, IMAI Hideyuki, MIYAKOSHI Masaaki

  The Transactions of the Institute of Electronics,Information and Communication Engineers. A 一般社団法人電子情報通信学会 J85-A (6) 730 - 734 0913-5707 2002/06 [Refereed][Not invited]
   

  一般に,分離核作用素に対する正則化一般逆行列は分離核とはならない.したがって,その算出には,Kronecker積及びvec作用素を用いた展開形を用いることになる.しかしながら,計算に要する記憶領域及び計算量双方の観点から,当該展開形を用いずに算出した方が効率が良いことは想像にかたくない.本論文では,分離核表現可能な正則化一般逆行列を提案することを目的とする.また,提案正則化一般逆行列の正則化効果についての定量的な評価も併せて行う.
• Choosing the Parameter of Image Restoration Filters by Modified Subspace Information Criterion

  TANAKA Akira, IMAI Hideyuki, MIYAKOSHI Masaaki

  IEICE Transactions on Fundamentals 一般社団法人電子情報通信学会 E85-A (5) 1104 - 1110 0916-8508 2002/05 [Refereed][Not invited]
   

  Practical image restoration filters usually include a parameter that controls regularizability, trade-off between fidelity of a restored image and smoothness of it, and so on. Many criteria for choosing such a parameter have been proposed. However, the relation between these criteria and the squared error of a restored image, which is usually used to evaluate the restoration performance, has not been theoretically substantiated. Sugiyama and Ogawa proposed the subspace information criterion (SIC) for model selection of supervised learning problems and showed that the SIC is an unbiased estimator of the expected squared error between the unknown model function and an estimated one. They also applied it to restoration of images. However, we need an unbiased estimator of the unknown original image to construct the criterion, so it can not be used for general situations. In this paper, we present a modified version of the SIC as a new criterion for choosing a parameter of image restoration filters. Some numerical examples are also shown to verify the efficacy of the proposed criterion.
• Choosing the Parameter of Regularized MP-Inverses by Modified SIC in Image Restoration Problems

  A. Tanaka, H. Imai, M. Miyakoshi

  International Journal of Computing Anticipatory Systems 12 166 - 178 2002 [Refereed][Not invited]
  
• Image Restoration by the Family of M-N-Norm Parametric Projection Filters

  A. Tanaka, H. Imai, M. Miyakoshi

  Electronics and Communications in Japan, Part 3 85 (11) 9 - 17 2002 [Refereed][Not invited]
  
• 線形最適画像復元フィルタに関する一考察

  田中章, 今井英幸, 宮腰政明

  第16回DSPシンポジウム講演論文集 53 - 58 2001/11 [Not refereed][Not invited]
  
• Image Restoration by the Family of M-N-Norm Parametric Projection Filters

  TANAKA Akira, IMAI Hideyuki, MIYAKOSHI Masaaki

  The Transactions of the Institute of Electronics,Information and Communication Engineers. A 一般社団法人電子情報通信学会 J84-A (8) 1063 - 1070 0913-5707 2001/08 [Refereed][Not invited]
   

  画像復元問題において, 未知の原画像に関する知識が利用可能である場合, その知識を復元過程に反映させることにより効果的な復元が可能となる. しかし, 射影フィルタ族やパラメトリック射影フィルタ族等のいわゆる統計的画像復元法では, 当該知識として, 原画像の属する範囲を線形部分空間に限定したり, 空間の計量を画像母集団の分散構造から決定する等の限られた情報以外は利用できない. 筆者らは, この問題を解決すべく画像の知識の反映をパレート最適という形で導入した信号適応型パラメトリック射影フィルタ族を提案している. 本論文では, 新たなアプローチとして, 画像の知識を空間の計量に反映可能なフィルタ族の構成を行う. また, 当該計量の構成例を示し, 数値例によりその効果を検証する.
• Kullback-Leibler information consistent estimation for censored data

  A Suzukawa, H Imai, Y Sato

  ANNALS OF THE INSTITUTE OF STATISTICAL MATHEMATICS 53 (2) 262 - 276 0020-3157 2001/06 [Refereed][Not invited]
   

  This paper is intended as an investigation of parametric estimation for the randomly right censored data. In parametric estimation, the Kullback-Leibler information is used as a measure of the divergence of a true distribution generating a data relative to a distribution in an assumed parametric model M. When the data is uncensored, maximum likelihood estimator (MLE) is a consistent estimator of minimizing the Kullback-Leibler information, even if the assumed model M does not contain the true distribution. We call this property minimum Kullback-Leibler information consistency (MKLI-consistency). However, the MLE obtained by maximizing the likelihood function based on the censored data is not MKLI-consistent. As an alternative to the MLE, Oakes (1986, Biometrics, 42, 177-182) proposed an estimator termed approximate maximum likelihood estimator (AMLE) due to its computational advantage and potential for robustness. We show MKLI-consistency and asymptotic normality of the AMLE under the misspecification of the parametric model. In a simulation study, we investigate mean square errors of these two estimators and an estimator which is obtained by treating a jackknife corrected Kaplan-Meier integral as the log-likelihood. On the basis of the simulation results and the asymptotic results, we discuss comparison among these estimators. We also derive information criteria for the MLE and the AMLE under censorship, and which call be used not only for selecting models but also for selecting estimation procedures.
• 信号適応型パラメトリック射影フィルタ族による画像復元

  田中章, 今井英幸, 宮腰政明

  電子情報通信学会論文誌A J83-A (8) 1011 - 1020 2000/08 [Refereed][Not invited]
  
• ノルム拘束に基づくパラメトリック射影フィルタ族の正則化

  田中章, 今井英幸, 宮腰政明

  電子情報通信学会論文誌A J83-A (6) 812 - 820 2000/06 [Refereed][Not invited]
  
• 2次形式不等式制約を用いた射影フィルタによる画像復元

  田中章, 今井英幸, 宮腰政明

  電子情報通信学会論文誌D-II J82-D-II (3) 415 - 421 1999/03 [Refereed][Not invited]
  
• Evaluation of a Stability on Weights of Fuzzy Analytic Hierarchy Process Using a Sensitivity Analysis

  OHNISHI Shin-ichi, IMAI Hideyuki, KAWAGUCHI Michiaki

  Journal of Japan Society for Fuzzy Theory and Systems 日本知能情報ファジィ学会 9 (1) 140 - 147 0915-647X 1997 

  Analytic Hierarchy Process (AHP) proposed by Saaty, T.L. in 1977 is a method of decision making. Natural human sense can be reflected in it. Fuzzy AHP which is expansion of it can show substitutive and complementary selection. Using AHP or Fuzzy AHP, we often cannot find the best alternative, because the weights of the alternatives are not much different from the others. Then it is very important to investigate which element of a comparison matrix has most large influence on the weight. It makes us know the key of interpretation and structure of the data.Two metods to evaluate them using sensitivity analysis have been proposed. However they have problems. The one changes the structure of the data, and the other is complex. And a method to evaluate for Fuzzy AHP has not been proposed. In this paper we propose a simple methods which does't change the data suructure in AHP and evaluation in Fuzzy AHP, and show an numerical experiment.
• 多重解像度解析を用いたディジタル画像の拡大

  田中章, 今井英幸, 宮腰政明, 伊達惇

  電子情報通信学会論文誌D-II J79-D-II (5) 819 - 825 1996/05 [Refereed][Not invited]
  
• General Solutions to a Generalized Transitive Coupling Problem in Fuzzy Structural Modeling

  MIYAKOSHI Masaaki, IMAI Hideyuki, WATANABE Takenori, DA-TE Tsutomu

  Journal of Japan Society for Fuzzy Theory and Systems 日本知能情報ファジィ学会 8 (2) 310 - 321 0915-647X 1996 

  Some methods of fuzzy structural modeling are proposed as applications of fuzzy theory to structural modeling.FISM/Fuzzy is such a modeling among them. In this modeling a trainsitive coupling problem yields in embedding process. For the problem, however, there are merely some algorithms to find particular solutions to the transitive coupling problem. An algorithm to generate general solutions has not been jet found. In this paper, generalizing the transitive coupling problem to a problem to solve simultaneous equations with two fuzzy eigen relation equations and a fuzzy relation inequality we present an algorithm to generate solutions to the generalized problem, and also it is clear that the presented algorithm consequently generates all solutions to the transitive coupling problem.

MISC

• 線形制約の一般解によるパラメトリック部分射影フィルタの解釈とアフィン制約付き復元問題への応用

  田中章, 今井英幸, 宮腰政明  電子情報通信学会論文誌  J89-A-  (8)  679  -681  2006  [Not refereed][Not invited]
  
• On formulations and solutions in linear image restoration problems

  A Tanaka, H Imai, M Miyakoshi  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E87A-  (8)  2144  -2151  2004/08  [Not refereed][Not invited]
   

  In terms of the formulation of the optimality, image restoration filters can be divided into two streams. One is formulated as an optimization problem in which the fidelity of a restored image is indirectly evaluated, and the other is formulated as an optimization problem based on a direct evaluation. Originally, the formulation of the optimality and the solutions derived from the formulation are identical each other. However in many studies adopting the former stream, an arbitrary choice of a solution without a mathematical ground passes unremarked. In this paper, we discuss the relation between the formulation of the optimality and the solution derived from the formulation from a mathematical point of view, and investigate the relation between a direct style formulation and an indirect one. Through these analyses, we show that the both formulations yield the identical filter in practical situations.
• Enhanced interval splitting and bounding for global optimization

  RW Mwangi, H Imai, Y Sato  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E87A-  (8)  2120  -2125  2004/08  [Not refereed][Not invited]
   

  In order to produce precise enclosures from a multidimensional interval vector, we introduce a sharp interval sub-dividing condition for optimization algorithms. By utilizing interval inclusion properties, we also enhance the sampling of an upper bound for effective use in the interval quadratic method. This has resulted in an improvement in the algorithm for the unconstrained optimization problem by Hansen in 1992.
• On formulations and solutions in linear image restoration problems

  A Tanaka, H Imai, M Miyakoshi  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E87A-  (8)  2144  -2151  2004/08  [Not refereed][Not invited]
   

  In terms of the formulation of the optimality, image restoration filters can be divided into two streams. One is formulated as an optimization problem in which the fidelity of a restored image is indirectly evaluated, and the other is formulated as an optimization problem based on a direct evaluation. Originally, the formulation of the optimality and the solutions derived from the formulation are identical each other. However in many studies adopting the former stream, an arbitrary choice of a solution without a mathematical ground passes unremarked. In this paper, we discuss the relation between the formulation of the optimality and the solution derived from the formulation from a mathematical point of view, and investigate the relation between a direct style formulation and an indirect one. Through these analyses, we show that the both formulations yield the identical filter in practical situations.
• On detecting interactions in Hayashi's second method of quantification

  H Imai, D Izawa, K Yoshida, Y Sato  MODELING DECISIONS FOR ARTIFICIAL INTELLIGENCE, PROCEEDINGS  3131-  205  -214  2004  [Not refereed][Not invited]
   

  The method of quantification were developed and investigated for the purpose of analyzing qualitative data. In the second method of quantification, the matter of interest is to discriminate the categories of the response variable. For that purpose, numerical scores of each categories are introduced so that the categories of the response variable can be discriminated as well as possible by those score. Since the total score is the sum of each category's score, the model is an additive model. Thus, if observations have a synergism, the method fails to grasp the structure. As a consequence, the response variable seems not to be discriminated by the method. In this paper, we propose an extension of Hayashi's second method of quantification by applying a fuzzy integral approach. To use the degree of decomposition of scores, we can include interactions between categories to the model.
• Upper limits for criteria for tests of dimensionality under elliptical populations

  K Yoshida, H Imai, Y Sato  COMMUNICATIONS IN STATISTICS-THEORY AND METHODS  33-  (11-12)  2799  -2816  2004  [Not refereed][Not invited]
   

  In this paper, tests of dimensionality in the multivariate analysis of variance model under elliptical populations are considered. We show that the upper bounds proposed by Yoshida et al. [Yoshida, K., Imai. H., Sato, Y. (2002). New criteria for tests of dimensionality under elliptical populations. J Japan Statist. Soc. 32:183-192] are upper limits for likelihood ratio (LR), Lawley-Hotelling (LH), and Bartlett-Nanda-Pillai (BNP) criteria under elliptical populations.
• A MINIMIZATION METHOD FOR COMPUTING PARAMETER BOUNDS IN AN INTERVAL VALUED LINEAR REGRESSION MODEL USING INTERVAL ANALYSIS(Theory and Applications)

  Mwangi Ronald Waweru, Imai Hideyuki, Sato Yoshiharu  Journal of the Japanese Society of Computational Statistics  17-  (1)  21  -31  2004  [Not refereed][Not invited]
   

  Interval data may appear in perturbed problems or when measured quanties are subject to error. Small errors in data also occur because of roundoff. We have used interval analysis to bound data values and their arithmetic solutions. An algorithm that uses the interval quadratic and Newton methods to enclose minimum values is used in a least-squares objective function. As a result, appropriate bounds for the coefficient of a regression model have been obtained. An example from cardiology in Billard and Diday (2000) is illustrated.
• Digital Image Enlargement Based on Kernel Component Estimation

  TANAKA A.  International Journal of Computing Anticipatory Systems  15, 97-108-  97  -108  2004  [Not refereed][Not invited]
  
• Parameter estimation and image restoration using the families of projection filters and parametric projection filters

  H Imai, YY Yuan, Y Sato  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E85A-  (8)  1966  -1969  2002/08  [Not refereed][Not invited]
   

  It is widely known that the family of projection filters includes the generalized inverse filter, and that the family of parametric projection filters includes parametric generalized projection filters. However, relations between the family of parametric projection filters and constrained least squares filters are not sufficiently clarified. In this paper, we consider relations between parameter estimation and image restoration by these families. As a result, we show that the restored image by the family of parametric projection filters is a maximum penalized likelihood estimator, and that it agrees with the restored image by constrained least squares filter under some suitable conditions.
• Parameter estimation and image restoration using the families of projection filters and parametric projection filters

  H Imai, YY Yuan, Y Sato  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E85A-  (8)  1966  -1969  2002/08  [Not refereed][Not invited]
   

  It is widely known that the family of projection filters includes the generalized inverse filter, and that the family of parametric projection filters includes parametric generalized projection filters. However, relations between the family of parametric projection filters and constrained least squares filters are not sufficiently clarified. In this paper, we consider relations between parameter estimation and image restoration by these families. As a result, we show that the restored image by the family of parametric projection filters is a maximum penalized likelihood estimator, and that it agrees with the restored image by constrained least squares filter under some suitable conditions.
• A formula for normalizing transformation of some statistics

  N Taneichi, Y Sekiya, A Suzukawa, H Imai  COMMUNICATIONS IN STATISTICS-THEORY AND METHODS  31-  (2)  163  -179  2002  [Not refereed][Not invited]
   

  On the basis of Konishi's study of a normalizing transformation (Konishi (1)), a concrete normalizing transformation is derived. Some applications of the proposed normalizing transformation are shown, and performance of the transformation in the applications is numerically investigated.
• Properties of the solution set of a fuzzy relation equation form the viewpoint of attainability (共著)

  Soft Computing  6(2), 87-91-  2002  [Not refereed][Not invited]
  
• Properties of the solution set of a fuzzy relation equation form the viewpoint of attainability (共著)

  Soft Computing  6(2), 87-91-  2002  [Not refereed][Not invited]
  
• A Study on the Convergence Property in Genetic Algorithms by Markov Chain

  ITO Takayuki, IMAI Hideyuki, MIYAKOSHI Masaaki  The Transactions of the Institute of Electronics,Information and Communication Engineers.  84-  (9)  1454  -1456  2001/09/01  

  本論文では, 斉次有限マルコフ連鎖を用いた遺伝的アルゴリズムの定式化に基づき, 遺伝的アルゴリズムの収束性において, 各遺伝的操作が果たす役割について考察する.
• D-3 Elliptical分布における多変量分散分析の次元検定について(多変量解析(5))(日本統計学会第69回大会記録)

  吉田 清隆, 今井 英幸, 佐藤 義治  日本統計学会誌  31-  (3)  409  -409  2001
• The period of powers of a fuzzy matrix

  H Imai, Y Okahara, M Miyakoshi  FUZZY SETS AND SYSTEMS  109-  (3)  405  -408  2000/02  [Not refereed][Not invited]
   

  A fuzzy matrix plays an important role in system models based on a fuzzy relation, and the performance of such a system depends on forming powers of a fuzzy matrix. The powers of a general fuzzy matrix either converge or oscillate with a finite period. Many properties about convergence have been investigated, for example, sufficient conditions for convergence are shown. Moreover, a lot of properties about convergence of powers of a transitive fuzzy matrix have been considered. However, few investigations about the period have been made. In this paper, we consider the period of the powers of a general fuzzy matrix by a graph theoretical viewpoint, and show conditions for convergence under the max-min composition. (C) 2000 Elsevier Science B.V. All rights reserved.
• The family of regularized parametric projection filters for digital image restoration

  H Imai, A Tanaka, M Miyakoshi  IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES  E82A-  (3)  527  -534  1999/03  [Not refereed][Not invited]
   

  Optimum filters for an image restoration are formed by a degradation operator, a covariance operator of original images, and one of noise. However, in a practical image restoration problem, the degradation operator and the covariance operators are estimated on the basis of empirical knowledge. Thus, it appears that they differ from the true ones. When we restore a degraded image by an optimum filter belonging to the family of Projection Filters and Parametric Projection Filters, it is shown that small deviations in the degradation operator and the covariance matrix can cause a large deviation in a restored image. In this paper, we propose new optimum filters based on the regularization method called the family of Regularized Projection Filters, and show that they are stable to deviations in operators. Moreover, some numerical examples follow to confirm that our description is valid.
• A-3 非線形回帰モデルにおけるパラメータ変換について(日本統計学会第67回大会記録 : 統計一般理論(5)回帰分析)

  桑原 宏和, 今井 英幸, 佐藤 義治  日本統計学会誌  29-  (3)  400  -400  1999
• Unattainable solutions of a fuzzy relation equation

  H Imai, K Kikuchi, M Miyakoshi  FUZZY SETS AND SYSTEMS  99-  (2)  193  -196  1998/10  [Not refereed][Not invited]
   

  In a fuzzy relation equation, the properties of the solution set have been investigated especially when the basic spaces are all finite sets. However, when the basic spaces are infinite, a few properties are known. In this paper, we use the concepts of attainability and unattainability of a solution to clarify some properties of the solution set and show the condition for existence of a partially attainable solution and an unattainable one. (C) 1998 Elsevier Science B.V. All rights reserved.
• A method of identifying influential data in fuzzy clustering

  H Imai, A Tanaka, M Miyakoshi  IEEE TRANSACTIONS ON FUZZY SYSTEMS  6-  (1)  90  -101  1998/02  [Not refereed][Not invited]
   

  In multivariate statistical methods, it is important to identify influential observations for a reasonable interpretation of the data structure, In this paper, we propose a method for identifying influential data in the fuzzy C-means (FCM) algorithm, To investigate such data, we consider a perturbation of the data points and evaluate the effect of a perturbation. As a perturbation, we consider two cases: one is the case in which the direction of a perturbation is specified and the other is the case in which the direction of a perturbation is not specified, By computing the change in the clustering result of FCM when given data points are slightly perturbed, we can look for data points that greatly affect the result, Also, we confirm an efficacy of the proposed method by numerical examples.
• A power approximation of the test of homogeneity for multinomial populations based on a normalizing transformation

  N Taneichi, Y Sekiya, H Imai  COMMUNICATIONS IN STATISTICS-THEORY AND METHODS  27-  (7)  1677  -1691  1998  [Not refereed][Not invited]
   

  Cressie and Read(1984) introduced multinomial goodness-of-fit statistics based on a, class of divergence measures I-a between discrete distributions. It was also introduced that a class of statistics R-a for the test of homogeneity for multinomial populations based on I-a (Read and Cressie(1988)). This class includes Pearson's Chi(2) statistic (when a = 1) and the log-likelihood ratio statistic (when a = 0). All R-a have the same chi-squared limiting null distribution. The power of the class is ordinary approximated front a, noncentral chi-squared distribution that is also the same for all a. Applying the power approximation theories for the multinomial goodness-of-fit test developed by Broffitt and Randles(1977) and Drost et al.(1989), Taneichi and Sekiya(1995) proposed three approximations to the power of R-a that vary with the statistic chosen: In this paper we propose a new approximation to the power of R-a. The new approximation is a normal approximation based on a normalizing transformation of the statistics. Moreover, the proposed approximation and the other approximations are numerically compared with the power directly calculated from the product multinomial model. As a result of comparison, the new approximation is found to be effective for the statistics when a less than or equal to 0.
• On restoration of overlapping images

  H Imai, Y Nakata, M Miyakoshi  IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS  E80D-  (12)  1190  -1194  1997/12  [Not refereed][Not invited]
   

  We consider the situation that plural degraded images are obtained, When no prior knowledge about original images are known, these images are individually restored by an optimum restoration filter, for example, by Wiener Filter or by Projection Filter. If correlations between original images are obtained, same restoration filters based on Wiener Filter or Projection Filter are proposed, In this paper, we deal with the case that some pixels or some gaits of original images overlap, and propose a restoration method using a formulae for overlapping. The method is based on Partial Projection Filter. Moreover, we confirm an efficacy of the proposed method by numerical examples.
• Some properties of minimal solutions for a fuzzy relation equation

  H Imai, M Miyakoshi, T Date  FUZZY SETS AND SYSTEMS  90-  (3)  335  -340  1997/09  [Not refereed][Not invited]
   

  Some properties of the solution set and minimal solutions of a fuzzy relation equation are considered. In this paper, we show the necessary and sufficient condition for existence of a minimal solution of a finite fuzzy relation equation defined on infinite index sets. (C) 1997 Elsevier Science B.V.
• The family of parametric projection filters and its properties for perturbation

  H Imai, A Tanaka, M Miyakoshi  IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS  E80D-  (8)  788  -794  1997/08  [Not refereed][Not invited]
   

  A lot of optimum filters have been proposed for an image restoration problem. Parametric filter, such as Parametric Wiener Filter, Parametric Projection Filter, or Parametric Partial Projection Filter, is often used because it requires to calculate a generalized inverse of one operator. These optimum filters are formed by a degradation operator, a covariance operator of noise, and one of original images. In practice, these operators are estimated based on empirical knowledge. Unfortunately it happens that such operators differ from the true ones. In this paper, we show the unified formulae of inducing them to clarify their common properties. Moreover, we investigate their properties for perturbation of a degradation operator. a covariance operator of noise. and one of original images. Some numerical examples follow to confirm that our description is valid.
• On Restoration for Overlapping Images

  NAKATA Yasuhisa, IMAI Hideyuki, MIYAKOSHI Masaaki  Proceedings of the IEICE General Conference  1997-  (2)  82  -82  1997/03/06  

  観測画像から原画像を推定する方法において、ただ1枚の観測画像のみではなく、観測画像が複数枚入手できる場合がある。そのような場合には、1枚の画像の復元に多くの観測画像を有効に利用することが復元精度の向上において重要となる。本文では、共通部分を持つ複数枚の画像を、部分射影フィルタを用いて復元する手法を提案する。
• 摂動法を用いた最適画像復元フィルタの評価(共著)

  電子情報通信学会論文誌  J80-D-II-  (3)  754  -761  1997  [Not refereed][Not invited]
  
• 摂動に対する射影フィルタ族の性質(共著)

  電子情報通信学会論文誌  J80-D-II-  (5)  1128  -1135  1997  [Not refereed][Not invited]
  
• Properties of the Family of Projection Filters for a Perturbation of Operators.

  IMAI H.  The Transactions of the Institute of Electronics, Information and Communication Engineers  J80-D-II-  (5)  1128  -1135  1997  [Not refereed][Not invited]
  
• Evaluation of Optimum Image Restoration Filters by Perturbation of a Degradation Operator

  The Transactions of the Institute of Electronics, Information and Communication Engineers  J80-D-II-  (3)  754  -761  1997  [Not refereed][Not invited]
  
• Evaluation of optimum image restoration filters by perturbation of a degradation operator

  Hideyuki Imai, Akira Tanaka, Masaaki Miyakoshi  Systems and Computers in Japan  28-  (8)  25  -32  1997  [Not refereed][Not invited]
   

  The image restoration problem may be considered in the framework of Hilbert space. In such a framework, the optimal restoration filter is made of a degradation operator and a covariance operator of noise. In practice, the degradation operator is determined by experimental knowledge. Unfortunately, it happens that such a degradation operator differs from the true one. Thus, the discrepancy between these operators has essential effects on the restored image. In this paper, confining ourselves to a real-valued finite-dimensional space, we present a method of evaluating how the optimal restoration filter is affected by perturbation of a degradation operator. Some numerical examples follow to confirm that our description is valid. © 1997 Scripta Technica, Inc.
• An Algebrac Structure of the Set of Solutions to a Transitive Coupling Problem in Fuzzy Structural Modeling

  MIYAKOSHI Masaaki, IMAI Hideyuki  Journal of Japan Society for Fuzzy Theory and Systems  8-  (4)  629  -638  1996/08/15  

  A transitive coupling problem discussed by Wakabayashi and Ouchi that arises in the embedding process of their method for fuzzy structural modeling. They also have considered the particular max and min solutions for the problem. The present paper aims at disclosing algebraic structure of the solution set for the problem based on a study of a generalized transitive coupling problem by the authors, whereby the followings are shown : ・the solution set is closed under scalar product, ・the solution set is closed under ∧-, ∨-operations, but not necessarily under ∨-, ∨-operations, ・if the solution set includes a solution which is different from the self-evident solution, the solution set has the carbinal number of continuum.Moreover, considering the relationship between the particular and the general solutions, the authors show that ・the min solutions decide whether the third inequality with α-product plays the role of a substaintial constrain or not, ・the min solutions are extremal in the solution set.
• A Note on Designing a Distributed System using OOD

  高桑 伸行, 今井 英幸, 宮腰 政明, 伊達 惇  全国大会講演論文集  51-  125  -126  1995/09/20  

  ソフトウェアシステムを設計するために、オブジェクト指向モデルが持つ優れた能力を用いた方法論が展開されてきた。ソフトウェアを部品の組立で構成できる反面、同一クラス内の複数のオブジェクトが存在するような分散システムでは、オブジェクト群の同期をとるといった時間的把握が困難である。本研究では、オブジェクト指向設計技法におけるオブジェクトの状態遷移を示す動的モデルの作成に際して、構造化設計技法であるJSD法の時間の取扱いを参考とした設計を行う。また、そのモデルとしてTime Petri Netの記述法を利用して書き表し、分散型電子会議支援システムに対して設計を行った結果を報告する。
• A Method for Image Enlargement Using Multiresolution Analysis

  TANAKA Akira, IMAI Hideyuki, MIYAKOSHI Masaaki, DA-TE Tsutomu  Proceedings of the IEICE General Conference  1995-  (2)  57  -57  1995/03/27  

  Mallatにより提案された多重解像度解析は,2乗可積分関数空間を解像度の異なる部分空間列により表現する.各部分空間は一つ解像度の低い部分空間と,その直交補空間の直和として表現され,Mallatはこれを直交ウェーブレットとスケール関数を用いて記述した.本研究は,この多重解像度解析により得られた各系列の相関に着目し,画像の拡大に応用する試みである.
• ファジィデータに対する射影追跡の適用(共著)

  日本ファジィ学会誌  7-  (6)  1239  -1246  1995  [Not refereed][Not invited]
  
• Exploratory Projection Pursuit for Fuzzy Data

  IMAI Hideyuki, SATO Yoshiharu  Journal of Japan Society for Fuzzy Theory and Systems  7-  (6)  1239  -1246  1995  [Not refereed][Not invited]
   

  Projection Pursuit is a computer-intensive method for multivariate statistical analysis which intends to clarify properties of multivariate data by projecting them onto lower-dimensional subspace.In practice, the projection pursuit is performed based on an algorithm to find a lower-dimensional subspace by maximizing an objective function called projection index which is a measure of interestingness of the projected data.Adapting this method for fuzzy data, it was proposed to use measures of fuzziness as projection index, and to find the subspace by minimizing the measures of fuzziness of projected data. So the subspace in which they are as crisp as possible is searched.But from the viewpoint of statistical analysis, important properties such as cluster or functional relationship are not considered.In this paper, we propose a method that includes both interestingness of data and the measures of fuzziness called hybrid index. And we confirm the efficiency of projection pursuit using this index by numerical experiments.
• A Study of the Inclusion Mapping Structure of Divergence-Convergence Boundary of Homogeneous Quadratic Transformation corresponding to Logistic Mapping

  吉川 毅, 今井 英幸, 宮腰 政明, 伊達 惇  全国大会講演論文集  49-  113  -114  1994/09/20  

  2次元同次二次変換はT.Date and M.Iriにより、不変式系の立場から10個の型に分類がなされている。本論文では良く知られるジュリア集合のうち、変換の反復により漸近的に原点に収束するような点の初期値の集合を収束領域、その上限を発散収束境界と呼び、その発散収束境界の形状について考察する。これは変換式によって多様な形状を示す場合があるが、同じ分類の型の中で複雑な形状が出現するパラメータと出現しないパラメータがある。今回は、そのような変換の例として方向に関する変換がロジスティック写像となるものを取り上げで考察した。
• Rsgion Segmentation of Color Images by Fuzzy Clustering

  宮達 安夫, 今井 英幸, 宮腰 政明, 伊達 惇  全国大会講演論文集  49-  321  -322  1994/09/20  

  画像情報を構造的に理解する上で、画像の領域分割は必要不可欠な手段であり、その手法は原画像の存在する原画像空間、あるいはその特徴量により作られる特徴空間を用いる場合の二つに分類される。本研究では特徴空間である色空間の上でクラスタリングをすることにより、画像の領域分割を行なう。従来の研究ではクラスタリング手法として主にkmeans法を用いているが、本研究ではより柔軟な分割が可能なファジィクラスタリング手法であるfuzzy cmeans(FCM)法を用いる。この手法の問題点としては初期分割の与え方により分割結果が相異なる場合があること、および他の領域に比べて小さな領域が存在するとき、この小領域をうまく分割できないことなどが可能であるファジィクラスタリングも考えられている。しかし、その手法も各クラスターの体積に制約を課す必要があり小さな領域は分割できない。そこでこの点を改善するために、重み乗数の性質を利用したファジィクラスタリングを用いることを提案する。この手法を用いると直前のステップで得られた情報を利用することができ、小さな領域に関しても分割が可能となる。
• Properties of a set of minimal solutions for a fuzzy relation equation

  Miyakoshi Masaaki, Imai Hideyuki, Da-te Tsutomu  Bulletin of the Faculty of Engineering, Hokkaido University  167-  1  -10  1994/01
• The fuzzy clustering algorithm based on the properites of the weighting exponent

  Imai Hideyuki, Miyazawa Yasuo, Miyakoshi Masaaki  Bulletin of the Faculty of Engineering, Hokkaido University  167-  11  -18  1994/01
• 対応分析における布置図の信頼性の視覚的表示法(共著)

  計算機統計学  7-  (2)  17  -23  1994  [Not refereed][Not invited]
  
• Graphical Representation of Reliability for Configuration of Correspondence Analysis

  Bulletin of The Conputational Statistics of Japan  7-  (2)  17  -23  1994  [Not refereed][Not invited]
  
• Evaluation of Effect on Configurations by Perturbation of The Data Points

  今井 英幸, 佐藤 順一, 伊達 惇  Bulletin of the Faculty of Engineering,Hokkaido University  (165)  p53  -59  1993/07
• 距離の重みを調整するアルゴリズムに基づくクラスタリング (日本計算機統計学会第6回大会報告)

  藤井 一郎, 今井 英幸, 佐藤 義治  計算機統計学  5-  (2)  181  -182  1992/12/18
• A Test for Redundancy of Some Sets of Variables

  今井 英幸, 伊達 惇  Bulletin of the Faculty of Engineering,Hokkaido University  (161)  p23  -27  1992/10
• 射影追跡法における感度分析(共著)

  計算機統計学  5-  (2)  101  -106  1992  [Not refereed][Not invited]
  
• Sensitivity Analysis for Exploratory Projection Pursuit.

  Imai Hideyuki, Sato Yoshiharu  Bulletin of the Computational Statistics of Japan  5-  (2)  101  -106  1992  [Not refereed][Not invited]
   

  Projection pursuit is a computer-intensive method for statistical analysis which intends to clarify property of multivariate data by projecting them onto a lower-dimensional subspace. In practice, the projection pursuit is performed based on an algorithm to find a lower-dimensional subspace by maximizing an objective function called projection index which is a measure of interestingness of the projected data. In this paper, we try to evaluate how the direction of projection determined in the process of the projection pursuit is affected by perturbation of the data points.

Books etc

• On detecting interactions in Hayashi's second method of quantification. 'jointly worked'

  Modeling Decisions for Artificial Intelligence 2004
• Algorthm based on alternative projections for a fuzzy measure identification problem 'jointly worked'

  Information Fusion in Data Mining, Springer 2003
• Algorthm based on alternative projections for a fuzzy measure identification problem 'jointly worked'

  Information Fusion in Data Mining, Springer 2003

Association Memberships

• 日本行動計量学会   日本統計学会   

Research Projects

• MARE: Recognition and Discovery Techniques for Giving Awarenesss

  Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research

  Date (from‐to) : 2019/04 -2024/03 

  Author : 工藤 峰一, 今井 英幸, 中村 篤祥

   

  今期は、不頻出事象の予測を抽象化したロングテール問題について基礎的検討を重ねた。ロングテールの特徴づけとして、極度にインバランス（サンプル数がクラスによって大きく異なる）であること、マルチラベル分類が多いこと、から、マルチラベル分類とインバランス問題の解決に注力した。さらに、実世界の不頻出事象に関連して、希少疾患予測の検討および独居老人の異変行動シミュレーションを行った。詳細を以下に示す。
  ・インバランスが顕著なマルチクラス問題を扱うため、2種類のクラス決定木（葉が一つクラスである決定木）を検討した。クラス集合対としてバランスをとるよう木を構成することでサンプル数の少ないクラスの検出精度を高めることに成功した。さらに、サンプルの少ないクラスが「次元の呪い」を受けやすいことを考慮して、特徴選択を少なくするようにクラス決定木を構成することで少数クラスのみならず全体の性能を向上させられることを示した。解釈可能性も大きく改善した。 ・希少疾患の診断において、１）医者の”希少疾患見過ごし”を防ぐために可視化手法を利用することを提案した、さらに、２）特徴選択により希少疾患の診断精度を向上させられることを確認した。 ・不頻出事象の予測に関して、極値論とサポートベクターマシンを利用した未知クラスの発見手法を提案し、これまでの手法よりよい性能を達成した。 ・独居老人の異変検知手法を開発する前段階として、異変を含む仮想住人の行動シミュレータを開発した。高齢者を実際のスマートホームに住まわせ、転倒などの異変が起きるのを待つわけにはいかないため、シミュレータは必須である。また、長期間に渡るセンサデータの採取や多様な生活行動を扱う上でも有効である。残念ながらこれまでの方式で複数の異変を適切に扱えるものがなかったため、確率モデルを利用したシミュレータを新たに開発した。
• Graph-based Information Theoretic Semi-Supervised Learning

  Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research

  Date (from‐to) : 2015/04 -2018/03 

  Author : Yoshida Tetsuya, IMAI Hideyuki

   

  In order to cope with increasing quantity and variety of data, it is important to develop information technology which enables effective use of domain knowledge. We have developed a graph-based information theoretic semi-supervised learning method. In the developed method, the relationship among data is represented as a graph based on mutual information, and domain knowledge is regarded as constraints and used for regularization. Under the framework of optimization learning, we have developed a semi-supervised learning algorith based on the representation matrix of the graph. The algorithm has been implemented as a prototype system, and experiments over the prototype system were conducted over several benchmark datasets. The results indicate the effectiveness of the developed learning method.
• eXtFS:Feature Selection and Exploration in extremly large multi-label classification problems

  Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research

  Date (from‐to) : 2015/04 -2018/03 

  Author : Kudo Mineichi

   

  We have dealt with "multi-label classification" problems where an object is assigned multiple labels. This study aimed at raising the classification performance and speeding up without degradation of performance. Our achievement is three of the following. First, we have pointed out the importance of the correlation between labels and showed several ways using it. Second, to keep a realistic processing time, we showed that the problem division of samples on the basis of their features or labels in some experimental results. Last, we pointed out the necessity of a special treatment on labels that appear rarely or have been forgotten to assign.
• Transfer Learning based on the structure of Feature Space

  Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research

  Date (from‐to) : 2012/04 -2016/03 

  Author : YOSHIDA Tetsuya, IMAI Hideyuki

   

  In order to cope with increasing quantity and variety of data, it is important to develop information technology which enables effective reuse of learned knowledge. We have developed a transfer learning method based on the structure of feature space. In the developed method, features are learned from the given data in the source domain, and a graph is constructed based on the learned features. Then, the method tries to preserve the structure of feature space as much as possible between the source domain and the target domain in transfer learning. Under the framework of optimization learning, we have developed a transfer learning algorithm based on multiplicative update rules. The algorithm has been implemented as a prototype system, and experiments over the prototype system were conducted over several benchmark datasets. The results of the experiments indicate the effectiveness of the developed transfer learning method.
• Classification of large scale qualitative data and visualization of their structure.

  Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research

  Date (from‐to) : 2011 -2013 

  Author : IMAI Hideyuki, KUDO Mineichi, TANAKA Akira

   

  To perform knowledge discovery from large data sets as signals from various sensors or texts accumulated in the web, it is required to classify into groups which have some common attribute as a pretreatment for detailed analysis even for such large and irregular data sets. Though such data sets contain a large number of items, it is often sufficient to use a small number of items for reasonable classification. We have studied mainly variable selection and regularization method for classification of qualitative data both from theoretical analysis and numerical experiments. These results have applied to the system for estimating human behavior using the data from the infrared sensors placed in the ceiling of the room.
• Resampling method for singular statistical model

  Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research

  Date (from‐to) : 2005 -2008 

  Author : IMAI Hideyuki, SATO Yoshiharu, KUDO Mineichi, TANEICHI Nonuhiro, MURAI Tetsuya, SAKURAI Hirohito

   

  観測されたデータがどのような集団に属しているかがはっきりしている場合には、どの方法を使って解析すればよいか、また、その手法がどのような性質を持っているかがよく判っている。一方、データの属する集団の形状によっては、今までに使われている手法がそのままでは使えない場合があることも判ってきた。そのような場合に、データを繰返し使って解析をする手法がどのくらい有効であるかを、数値実験と、理論の両面から考察した。
• 波面記述法による3次元空間のモデリング

  日本学術振興会：科学研究費助成事業

  Date (from‐to) : 2004 -2005 

  Author : 坂本 雄児, 今井 英幸

   

  1,研究概要 3次元映像をデジタルデータとして記録する方法として,波面を記録する波面記述法の基礎理論を確立する。この方法は,従来法に比べ格段の情報量削減の可能性があるが,実際的なアルゴリズムやその座標変換等の基礎理論において明らかになっていない点が多くあり,これを研究する。 2.平成17年度の研究状況 (1)波面記述法の有効性を検証する目的で,平成16年度に引続きホログラフィックディスプレイ装置を開発した。これを用いて,波面記述法からホログラムデータへの変換および光学再生実験を行い,立体像の表示に成功した。この実験を通して波面記述法とホログラフィックディスプレイの親和性の高さを確認した。 (2)多眼画像データなどの3次元画像入力装置で得られたデータの波面記録法への変換方法の検討を行い,変換アルゴリズムを明らかにした。さらに、上記(1)のシステムにより、立体像表示実験を行い、その有効性を明らかにした。 (3)この研究過程で,平成16年度に円筒形の記録面を用いた記録法が波面記述法に適していることが明らかになったが,本年度は,円筒形記録面での高速変換アルゴリズムの検討を行い,実用化の可能性のある非常に高速なアルゴリズムを明らかにした。
• Development of Classification Systems for Large-Scale Pattern Recognition Problems and Its Application

  Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research

  Date (from‐to) : 2002 -2005 

  Author : KUDO Mineichi, MURAI Tetsyua, IMAI Hideyuki, TOYAMA Jun, TENMOTO Hiroshi, HAYASHI Hiroki

   

  In this study, we classified the scaling problems of pattern recognition tasks into three of the following. The results are shown below, respectively. 1.Scaling problem as for data number : We have changed the problem setting from the problem to achieve as high (predictive) performance as possible to the problem to attain as least computational time as possible keeping no large damage in performance. Especially, 1)a fast k-nearest neighbor was invented, and 2)a one-pass algorithm for prototype acquisition was discussed, though they are still under progress. 2.Scaling problem as for dimensionality : We concentrated on "feature selection" which removes low-informative features but keeps the classification performance. Solving this problem usually require a combinatorial examination, so many sub-optimal approaches have been proposed so far. However, they are not satisfactory in time for large-scale problems in the number of original features. In this study, we focused on "classifier-independent feature selection" in which only garbage (completely no informative for classification) features are removed. Usual algorithms do "classifier-dependent feature selection." We showed that the former can be made faster than the latter, and a two-stage selection scheme (CIFS+CSFS) is efficient for feature selection. 3.Scaling problem as for cass number : As seen in Kanji-Character recognition, we sometimes have to deal with many categories. The difficulty of classification task increases as the class number increases. We proposed a way to analyze the relationship among classes and grouped some hard-to-classify classes into one "super class." In addition, we investigated a visualization method using a graph representation and a tree representation for making clear the relationship. In the decision tree, we confirmed that the total performance can be increased with feature selection. Last, we suggested that such super-classes and subclasses analysis is useful to capture the individual classification problems.
• Construction of Regularized Parametric Projection Filters

  Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research

  Date (from‐to) : 2001 -2002 

  Author : MIYAKOSHI Masaaki, IMAI Hideyuki, TANAKA Akira, KAWAGUCHI Mayuka

   

  Functional Analysis-Theoretical developments of Image Restoration Problems carried on, and many filter-families such as Projection Filter- and Parametric Projection Filter-families were proposed. In practice, since the fidelity of restored images to original images depends on the stability of used restoration filters, it is very important to strengthen their robustness. The study is to introduce a new filter-family that is robust against the perturbation of degraded processes for the parametric projection filter-family with less calculations and high fidelity. In the two-years term of project, we proposed the following three restoration filter-families and confirmed their effectiveness by numerical simulations : Regularized Parametric Projection Filter-Family, M-N-Norm Parametric Projection Filter-Family, and Signal Adaptive Parametric Projection Filter-Family. Moreover, we introduced a criterion to choose perturbation, parametric and regularized parameters. The criterion makes us possible to choose the optimal, parameter of Parametric Projection Filter-family in a sense, and it results in the higher fidelity of restored images.
• Studies of Possibilities and Limits of Statistical Prediction

  Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research

  Date (from‐to) : 1999 -2002 

  Author : SATO Yoshiharu, IMAI Hideyuki, MURAI Tetsuya, MIZUTA Masahiro, TANIEICHI Nobuhiro, BABA Yasumasa

   

  The aim of this project was to study out the possibilities and limits of statistical prediction in the practical problems. There are two situations in the statistical prediction. One is an interpolation and the other is an extrapolation. The typical concept of the former, interpolation, is a theory of regression, namely, the nonlinear, semi-parametric and nonparametric regression. Moreover, the several modeling techniques by which the reasonable and useful information are extracted have been included. On these methods, we get excellent results, for instance, "growth curve model with hierarchcal within-individual design matrices", "echelon analysis of the relationship between population and land cover patterns based on remote sensing data", "an algorithm with projection pursuit for sliced inverse regression model", "Kullback-Leibler information consistent estimator for censored data" and so on. On the other hand, the latter concept, extrapolation, is familiar to weather or a business or a stock price forecasting. The typical method is a time series analysis. Although, there still remain many dfficulties, in this project, many useful results are proposed, especially on the state space modeling a relaxation of error distribution and several filtering theory for the noise reduction. For example, "computational implementations of nonlinear non-Gaussian prediction and filtering formulas", "a combining forecast method using a probabilistic neural network", "forecasting structural changes in a state-space model", "extension of the MDL criterion by shrinkage method" and so on.
• Nonlinear Data Analysis with Algebraic Curve Fitting

  Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research

  Date (from‐to) : 1997 -2000 

  Author : MIZUTA Masahiro, SUZUKAWA Akio, MURAI Tetuya, SATO Yoshiharu, IMAI Hideyuki, MINAMI Hiroyuki

   

  The aim of the research is to construct a series of methods for nonlinear data analysis. Data analysis methods with algebraic curve fitting are fundamental methods in the research. We had studied algebraic curve fitting at first. We developed an algorithm to derive the exact distance between a point and algebraic curve. A method for finding an algebraic curve structures in data is proposed. According to the progress of the research, we also study Generalized Principal Components Analysis and Principal Curves. For related data analysis methods, Sliced Inverse Regression, Functional Data Analysis and Principal Points are studied. Sliced Inverse Regression with Projection Pursuit is proposed and numerical examples show the advantages of the proposed method. A method of nonlinear functional regression analysis is developed.
• Development of Visual Programming Environment for Exploratory Data Analysis

  Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research

  Date (from‐to) : 1995 -1997 

  Author : SATO Yoshiharu, MINAMI Hiroyuki, IMAI Hideyuki, SHIOYA Hiroyuki, MURAI Tetsuya, MIZUTA Masahiro

   

  In computer science, "Visualization" is one of the most attractive keywords. "Visualization" is useful not only to reveal data but to deal with analysis processes. Statisticians are often forced to apply many methods to the same data in their analysis since they wants to find the most reasonable result. A lot of results they made, however, may perplex them. If they can view an analysis flow (made of some statistical methods) on the graphical display, they can easily find the relations between data and applied methods. It is called "Visualization of Execution." As we described, "Visualization" has much power then we try to build "Visual Environment, " which synthesizes both "Visualization" merits. In short, our target is to build the system which can view, display data and construct analysis flows and execute them in the sameenvironment. In the system, if we can add and execute some statistical methods in an analysis flow, we can easily change some methods or apply another procedure to the original data or intermediate ones. We had already developed the prototypes and now improve it on the workstation with Tcl/Tk environment for its interoperability. In addition, this idea is also effective for novices in statistics but they have less knowledge on statistics then may make a meaningless statistical flow. We have already recognized this problem and tried to add a supporting module which makes use of technique on knowledge engineering. Especially, in the process of the research project, we have investigated the hard/software for the visual programming environment and have discussed the critical issues for the practical use of the environment
• 射影追跡法を用いたデータ解析の安定性について

  日本学術振興会：科学研究費助成事業

  Date (from‐to) : 1994 -1994 

  Author : 今井 英幸

   

  射影追跡法においては射影指標の性質を明らかにすることが重要である。本研究においては多項式指標に属している指標で、現在広く用いられているFriedmanによる指標とHallによる指標について感度分析法を応用した手法によって外れ値に対する安定性の観点から比較を行い、Hallの指標を用いた射影追跡方がより ロバスト(頑健)であることをいくつかの数値例によって示した。 データの平滑化手法として重要な射影追跡回帰については、従来、平滑化手法として用いられている移動平均法や重み付き局所回帰(LOWESS)などと比較するため、これらの従来手法の安定性について調べた。外れ値を含んでいるデータについて平滑化を行う手法と、平滑化に対して影響の大きい外れ値を摘出してから再度平滑化を行う手法の比較についても考察を行った。 また、言語変数データのように値そのものに曖昧さを含んでいるデータに対する射影追跡法の適用について、従来用いられている射影指標がデータ構造の曖昧さだけに注目していたことに対し、本研究においてはデータ構造の曖昧さと興味深さの双方を考慮にいれた指標を用いた射影指標によって射影追跡法を適用する手法を考案し、いくつかの具体例について実際に解析を行うことによってその有効性を確認した。 外れ値の摘出を含めたソフトウェアについては、一般の射影追跡法に対するグラフィック表示を行うところまで開発を行った。今後、射影追跡回帰および射影追跡密度推定のソフトウエアを開発する予定である。
• 非線形回帰分析
• Nonlinear Regression Analysis