Kenichi Satoh
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Kenichi Satoh has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
- Aya Goto
Fukushima Medical University
1 shared publication - Claudia Kimie Suemoto
Division of Geriatrics, University of Sao Paulo Medical School, Brazil
1 shared publication - Gen Ouchi
Department of Emergency and Critical Care Medicine, Graduate School of Medicine, University of the Ryukyus, Ginowan, Japan
1 shared publication - Katsuya Takemura
Department of Emergency and Critical Care Medicine, Graduate School of Medicine, University of the Ryukyus, Ginowan, Japan
1 shared publication - Kazuaki Uchida
National Center for Geriatrics and Gerontology, Obu, Aichi Japan
1 shared publication Kazuhiko Nakaharai
The Jikei University School of Medicine, 3-25-8, Nishi-Shimbashi, Minato-ku, Tokyo, 105-8461, Japan
1 shared publication
Latest Publications
- Open Access
- 31 August 2026
- Published by The Korean Society of Critical Care Medicine in Acute and Critical Care
- Vol. 41 (3) , 550-563
Minimum observation time to determine ventilator liberation success and its associated factors in tracheostomized patients in intensive and high-care units in Japan: a single-center retrospective study
Background: Time-based criteria for mechanical ventilation (MV) liberation in tracheostomized patients undergoing T-tube trials with ventilator discontinuation are inconsistent. We aimed to define the minimum post-liberation observation time and identify factors associated with successful liberation.Methods: This single-center retrospective cohort study included adults with tracheostomy who received MV in the intensive care unit and high care unit of a Japanese hospital from November 14, 2019, to August 31, 2024. We measured maximum endurance time (MET) without MV, MV liberation success, and the shortest MET when success first exceeded 80% (mini-MET). Time to mini-MET was assessed by Kaplan-Meier median survival time (MST), and competing risk analysis treating in-hospital death as a competing event was performed. Predictors were screened by univariate C-index, and survival tree cutoffs were examined using logistic regression sensitivity analyses.Results: Among 160 patients, MV liberation success increased with MET from 65% to 94% at 168 hours. The MST to achieve the mini-MET, which was identified as 4 hours, was 7 days (95% CI, 5–10 days). Mechanical power normalized by dynamic compliance (MPCdyn) ≤454.72 J/min and albumin >2.3 g/dl were associated with success. MST was 3 days when both were satisfied. In a covariate-adjusted model, low MPCdyn plus high albumin remained associated with success. Replacing MPCdyn with driving pressure yielded similar findings.Conclusions: Four hours may be a practical minimum post-liberation observation time, whereas 48 hours may be a conservative option. Stratification by MPCdyn or driving pressure plus albumin can support consistent MV liberation decisions after tracheostomy.
- 31 August 2026
- Published by The Korean Society of Critical Care Medicine in Acute and Critical Care
- Vol. 41 (3) , 550-563
Open AccessMinimum observation time to determine ventilator liberation success and its associated factors in tracheostomized patients in intensive and high-care units in Japan: a single-center retrospective study
Background: Time-based criteria for mechanical ventilation (MV) liberation in tracheostomized patients undergoing T-tube trials with ventilator discontinuation are inconsistent. We aimed to define the minimum post-liberation observation time and identify factors associated with successful liberation.Methods: This single-center retrospective cohort study included adults with tracheostomy who received MV in the intensive care unit and high care unit of a Japanese hospital from November 14, 2019, to August 31, 2024. We measured maximum endurance time (MET) without MV, MV liberation success, and the shortest MET when success first exceeded 80% (mini-MET). Time to mini-MET was assessed by Kaplan-Meier median survival time (MST), and competing risk analysis treating in-hospital death as a competing event was performed. Predictors were screened by univariate C-index, and survival tree cutoffs were examined using logistic regression sensitivity analyses.Results: Among 160 patients, MV liberation success increased with MET from 65% to 94% at 168 hours. The MST to achieve the mini-MET, which was identified as 4 hours, was 7 days (95% CI, 5–10 days). Mechanical power normalized by dynamic compliance (MPCdyn) ≤454.72 J/min and albumin >2.3 g/dl were associated with success. MST was 3 days when both were satisfied. In a covariate-adjusted model, low MPCdyn plus high albumin remained associated with success. Replacing MPCdyn with driving pressure yielded similar findings.Conclusions: Four hours may be a practical minimum post-liberation observation time, whereas 48 hours may be a conservative option. Stratification by MPCdyn or driving pressure plus albumin can support consistent MV liberation decisions after tracheostomy.
- Open Access
- 25 August 2026
- Published by Springer Nature in Japanese Journal of Statistics and Data Science
- p. 1-26
Applying non-negative matrix factorization with covariates to label matrix for classification
Non-negative matrix factorization (NMF) is widely used for representation learning and interpretable analysis in high-dimensional nonnegative data, but standard formulations are unsupervised and do not directly address classification. Existing supervised extensions typically incorporate labels only through penalties or graph constraints and often require an external classifier. We propose NMF-LAB (Non-negative Matrix Factorization for Label Matrix), a classification framework that directly factorizes the label matrix Y while treating covariates A as explanatory variables. By modeling class labels as the target of factorization, NMF-LAB yields a direct probabilistic mapping from covariates to labels without an additional classifier. The proposed framework gives rise to two complementary designs. NMF-LAB Direct emphasizes feature-level interpretability through its nonnegative and additive structure, whereas NMF-LAB Kernel incorporates nonlinear covariates to improve predictive performance and generalization. Kernel-based similarities are integrated via Gaussian-kernel covariates, and the kernel design scales to large datasets through Nyström approximation. Extensive experiments on diverse datasets, ranging from small- and medium-scale benchmarks to the large-scale MNIST dataset, demonstrate the interpretability of the Direct design through sparse, nonnegative coefficient estimation, and the competitive predictive accuracy of the Kernel design. Overall, NMF-LAB provides a unified, probabilistic, and scalable framework for classification based on nonnegative matrix factorization.
- 25 August 2026
- Published by Springer Nature in Japanese Journal of Statistics and Data Science
- p. 1-26
Open AccessApplying non-negative matrix factorization with covariates to label matrix for classification
Non-negative matrix factorization (NMF) is widely used for representation learning and interpretable analysis in high-dimensional nonnegative data, but standard formulations are unsupervised and do not directly address classification. Existing supervised extensions typically incorporate labels only through penalties or graph constraints and often require an external classifier. We propose NMF-LAB (Non-negative Matrix Factorization for Label Matrix), a classification framework that directly factorizes the label matrix Y while treating covariates A as explanatory variables. By modeling class labels as the target of factorization, NMF-LAB yields a direct probabilistic mapping from covariates to labels without an additional classifier. The proposed framework gives rise to two complementary designs. NMF-LAB Direct emphasizes feature-level interpretability through its nonnegative and additive structure, whereas NMF-LAB Kernel incorporates nonlinear covariates to improve predictive performance and generalization. Kernel-based similarities are integrated via Gaussian-kernel covariates, and the kernel design scales to large datasets through Nyström approximation. Extensive experiments on diverse datasets, ranging from small- and medium-scale benchmarks to the large-scale MNIST dataset, demonstrate the interpretability of the Direct design through sparse, nonnegative coefficient estimation, and the competitive predictive accuracy of the Kernel design. Overall, NMF-LAB provides a unified, probabilistic, and scalable framework for classification based on nonnegative matrix factorization.
- Open Access
- 3 July 2026
- Published by Taylor & Francis in American Journal of Mathematical and Management Sciences
- Vol. 45 (3) , 243-263
Applying Non-Negative Matrix Factorization with Covariates to the Longitudinal Data as Growth Curve Model
1. Non-negative Matrix Factorization (NMF) was proposed by Lee and Seung (1999) to approximate an observed matrix consisting of non-negative elements as the product of a basis and coefficient matri...
- 3 July 2026
- Published by Taylor & Francis in American Journal of Mathematical and Management Sciences
- Vol. 45 (3) , 243-263
Open AccessApplying Non-Negative Matrix Factorization with Covariates to the Longitudinal Data as Growth Curve Model
1. Non-negative Matrix Factorization (NMF) was proposed by Lee and Seung (1999) to approximate an observed matrix consisting of non-negative elements as the product of a basis and coefficient matri...
- 15 April 2026
- Published by The R Foundation
nmfkc: Non-Negative Matrix Factorization with Kernel Covariates
Performs Non-negative Matrix Factorization (NMF) with Kernel Covariates. Given an observation matrix and kernel covariates, it optimizes both a basis matrix and a parameter matrix. Notably, if the kernel matrix is an identity matrix, the method simplifies to standard NMF. Also provides NMF with Random Effects (NMF-RE) via nmfre(), which estimates a mixed-effects model combining covariate-driven scores with unit-specific random effects together with wild bootstrap inference, and NMF-based Structural Equation Modeling (NMF-SEM) via nmf.sem(), which fits a two-block input-output model for blind source separation and path analysis. References: Satoh (2025) ; Satoh (2025) ; Satoh (2025) ; Satoh (2026) ; Satoh (2026) .
- 15 April 2026
- Published by The R Foundation
nmfkc: Non-Negative Matrix Factorization with Kernel Covariates
Performs Non-negative Matrix Factorization (NMF) with Kernel Covariates. Given an observation matrix and kernel covariates, it optimizes both a basis matrix and a parameter matrix. Notably, if the kernel matrix is an identity matrix, the method simplifies to standard NMF. Also provides NMF with Random Effects (NMF-RE) via nmfre(), which estimates a mixed-effects model combining covariate-driven scores with unit-specific random effects together with wild bootstrap inference, and NMF-based Structural Equation Modeling (NMF-SEM) via nmf.sem(), which fits a two-block input-output model for blind source separation and path analysis. References: Satoh (2025) ; Satoh (2025) ; Satoh (2025) ; Satoh (2026) ; Satoh (2026) .
- Open Access
Applying non-negative matrix factorization with covariates to structural equation modeling for blind input-output analysis
Structural equation modeling (SEM) describes directed dependence and feedback, whereas non-negative matrix factorization (NMF) provides interpretable, parts-based representations for non-negative data. We propose NMF-SEM, a unified non-negative framework that embeds NMF within a simultaneous-equation structure, enabling latent feedback loops and a reduced-form input-output mapping when intermediate flows are unobserved. The mapping separates direct effects from cumulative propagation effects and summarizes reinforcement using an amplification ratio. We develop regularized multiplicative-update estimation with orthogonality and sparsity penalties, and introduce structural evaluation metrics for input-output fidelity, second-moment (covariance-like) agreement, and feedback strength. Applications show that NMF-SEM recovers the classical three-factor structure in the Holzinger-Swineford data, identifies climate- and pollutant-driven mortality pathways with negligible feedback in the Los Angeles system, and separates deprivation, general morbidity, and deaths-of-despair components with weak feedback in Mississippi health outcomes.
Open AccessApplying non-negative matrix factorization with covariates to structural equation modeling for blind input-output analysis
Structural equation modeling (SEM) describes directed dependence and feedback, whereas non-negative matrix factorization (NMF) provides interpretable, parts-based representations for non-negative data. We propose NMF-SEM, a unified non-negative framework that embeds NMF within a simultaneous-equation structure, enabling latent feedback loops and a reduced-form input-output mapping when intermediate flows are unobserved. The mapping separates direct effects from cumulative propagation effects and summarizes reinforcement using an amplification ratio. We develop regularized multiplicative-update estimation with orthogonality and sparsity penalties, and introduce structural evaluation metrics for input-output fidelity, second-moment (covariance-like) agreement, and feedback strength. Applications show that NMF-SEM recovers the classical three-factor structure in the Holzinger-Swineford data, identifies climate- and pollutant-driven mortality pathways with negligible feedback in the Los Angeles system, and separates deprivation, general morbidity, and deaths-of-despair components with weak feedback in Mississippi health outcomes.
- Open Access
Applying non-negative matrix factorization with covariates to label matrix for classification
Non-negative matrix factorization (NMF) is widely used for dimensionality reduction and interpretable analysis, but standard formulations are unsupervised and cannot directly exploit class labels. Existing supervised or semi-supervised extensions usually incorporate labels only via penalties or graph constraints, still requiring an external classifier. We propose \textit{NMF-LAB} (Non-negative Matrix Factorization for Label Matrix), which redefines classification as the inverse problem of non-negative matrix tri-factorization (tri-NMF). Unlike joint NMF methods, which reconstruct both features and labels, NMF-LAB directly factorizes the label matrix $Y$ as the observation, while covariates $A$ are treated as given explanatory variables. This yields a direct probabilistic mapping from covariates to labels, distinguishing our method from label-matrix factorization approaches that mainly model label correlations or impute missing labels. Our inversion offers two key advantages: (i) class-membership probabilities are obtained directly from the factorization without a separate classifier, and (ii) covariates, including kernel-based similarities, can be seamlessly integrated to generalize predictions to unseen samples. In addition, unlabeled data can be encoded as uniform distributions, supporting semi-supervised learning. Experiments on diverse datasets, from small-scale benchmarks to the large-scale MNIST dataset, demonstrate that NMF-LAB achieves competitive predictive accuracy, robustness to noisy or incomplete labels, and scalability to high-dimensional problems, while preserving interpretability. By unifying regression and classification within the tri-NMF framework, NMF-LAB provides a novel, probabilistic, and scalable approach to modern classification tasks.Open AccessApplying non-negative matrix factorization with covariates to label matrix for classification
Non-negative matrix factorization (NMF) is widely used for dimensionality reduction and interpretable analysis, but standard formulations are unsupervised and cannot directly exploit class labels. Existing supervised or semi-supervised extensions usually incorporate labels only via penalties or graph constraints, still requiring an external classifier. We propose \textit{NMF-LAB} (Non-negative Matrix Factorization for Label Matrix), which redefines classification as the inverse problem of non-negative matrix tri-factorization (tri-NMF). Unlike joint NMF methods, which reconstruct both features and labels, NMF-LAB directly factorizes the label matrix $Y$ as the observation, while covariates $A$ are treated as given explanatory variables. This yields a direct probabilistic mapping from covariates to labels, distinguishing our method from label-matrix factorization approaches that mainly model label correlations or impute missing labels. Our inversion offers two key advantages: (i) class-membership probabilities are obtained directly from the factorization without a separate classifier, and (ii) covariates, including kernel-based similarities, can be seamlessly integrated to generalize predictions to unseen samples. In addition, unlabeled data can be encoded as uniform distributions, supporting semi-supervised learning. Experiments on diverse datasets, from small-scale benchmarks to the large-scale MNIST dataset, demonstrate that NMF-LAB achieves competitive predictive accuracy, robustness to noisy or incomplete labels, and scalability to high-dimensional problems, while preserving interpretability. By unifying regression and classification within the tri-NMF framework, NMF-LAB provides a novel, probabilistic, and scalable approach to modern classification tasks. - Open Access
- 29 August 2025
- Published by Springer Nature in Japanese Journal of Statistics and Data Science
- Vol. 9 (1) , 79-97
Applying non-negative matrix factorization with covariates to multivariate time series data as a vector autoregression model
We propose a novel framework for analyzing multivariate time series (MTS) data by integrating non-negative matrix factorization (NMF) with vector autoregression (VAR). Termed NMF-VAR, this method models the coefficient matrix of NMF as a VAR process, enabling simultaneous extraction of latent components and temporal dependencies. Unlike standard VAR, which struggles with high dimensionality and lacks clarity, our method introduces a low-rank latent structure that reduces the number of parameters while retaining explanatory power. The proposed framework generalizes the standard VAR model to high-dimensional non-negative data, including the standard VAR as a special case. We formulate the estimation as a constrained optimization problem and present multiplicative update rules for NMF based on existing tri-factorization techniques. We evaluate the method on three real-world datasets: quarterly first-differenced macroeconomic indicators of Canada, monthly international airline passenger volumes, and daily COVID-19 infection counts across Japanese prefectures. The results demonstrate that NMF–VAR effectively captures meaningful patterns such as economic cycles, seasonal travel behavior, and regional epidemic trends. Moreover, the method yields a significant reduction in regression parameters, improving both scalability and model transparency. Overall, NMF–VAR provides an efficient and insightful tool for analyzing high-dimensional and large-scale time series data.
- 29 August 2025
- Published by Springer Nature in Japanese Journal of Statistics and Data Science
- Vol. 9 (1) , 79-97
Open AccessApplying non-negative matrix factorization with covariates to multivariate time series data as a vector autoregression model
We propose a novel framework for analyzing multivariate time series (MTS) data by integrating non-negative matrix factorization (NMF) with vector autoregression (VAR). Termed NMF-VAR, this method models the coefficient matrix of NMF as a VAR process, enabling simultaneous extraction of latent components and temporal dependencies. Unlike standard VAR, which struggles with high dimensionality and lacks clarity, our method introduces a low-rank latent structure that reduces the number of parameters while retaining explanatory power. The proposed framework generalizes the standard VAR model to high-dimensional non-negative data, including the standard VAR as a special case. We formulate the estimation as a constrained optimization problem and present multiplicative update rules for NMF based on existing tri-factorization techniques. We evaluate the method on three real-world datasets: quarterly first-differenced macroeconomic indicators of Canada, monthly international airline passenger volumes, and daily COVID-19 infection counts across Japanese prefectures. The results demonstrate that NMF–VAR effectively captures meaningful patterns such as economic cycles, seasonal travel behavior, and regional epidemic trends. Moreover, the method yields a significant reduction in regression parameters, improving both scalability and model transparency. Overall, NMF–VAR provides an efficient and insightful tool for analyzing high-dimensional and large-scale time series data.
- Open Access
- 28 October 2023
- Published by Elsevier in The Lancet Regional Health - Western Pacific
- Vol. 40, 100946
Two pills and women’s health in Japan
- 28 October 2023
- Published by Elsevier in The Lancet Regional Health - Western Pacific
- Vol. 40, 100946
Open AccessTwo pills and women’s health in Japan
- Open Access
- 15 August 2022
- Published by MDPI AG in International Journal of Environmental Research and Public Health
- Vol. 19 (16) , 10094
Factors Associated with Behavioral and Psychological Symptoms of Dementia during COVID-19
(1) Background: Preventive measures to control the spread of COVID-19 are essential, but they often cause social isolation and diminish the physical and mental health of older adults. In cognitively impaired individuals, the pandemic has worsened behavioral and psychological symptoms of dementia (BPSD). Here, we explored the factors contributing to the worsening of BPSD during the COVID-19 pandemic. (2) Methods: Potential patients were identified at a memory clinic in Japan between June 2017 and June 2021. Eligible patients had a diagnosis of mild cognitive impairment (MCI) or dementia during the study period. The outcome was BPSD, as assessed by using the Dementia Behavioral Disorders Scale. Information on patients’ lifestyle habits and use of care services was obtained for use as primary explanatory variables; multiple regression analysis was performed to examine the relationship between BPSD and care services use or lifestyle habits. The model was adjusted for sociodemographic factors, and the interaction terms of the pandemic period with lifestyle and service use were included to evaluate the effects of COVID-19. (3) Results: We identified 977 participants with MCI and 1380 with dementia (MCI group: 69.8% age 75 years or older, 54.2% female; dementia group: 79.8% age 75 years or older, 64.8% female). After adjustment for possible confounders, significantly worse BPSD was demonstrated in those who used daycare services during COVID-19 (both MCI and dementia patients; p = 0.007 and p = 0.025 respectively) and in those with poor nutritional function (dementia patients; p = 0.040). (4) Conclusions and Implications: During COVID-19, poor nutritional status and use of daycare services were associated with BPSD in those with cognitive decline. These findings indicate the need to fully examine the quantity and quality of care services for people with cognitive decline during emergencies and to continue to provide effective services.
- 15 August 2022
- Published by MDPI AG in International Journal of Environmental Research and Public Health
- Vol. 19 (16) , 10094
Open AccessFactors Associated with Behavioral and Psychological Symptoms of Dementia during COVID-19
(1) Background: Preventive measures to control the spread of COVID-19 are essential, but they often cause social isolation and diminish the physical and mental health of older adults. In cognitively impaired individuals, the pandemic has worsened behavioral and psychological symptoms of dementia (BPSD). Here, we explored the factors contributing to the worsening of BPSD during the COVID-19 pandemic. (2) Methods: Potential patients were identified at a memory clinic in Japan between June 2017 and June 2021. Eligible patients had a diagnosis of mild cognitive impairment (MCI) or dementia during the study period. The outcome was BPSD, as assessed by using the Dementia Behavioral Disorders Scale. Information on patients’ lifestyle habits and use of care services was obtained for use as primary explanatory variables; multiple regression analysis was performed to examine the relationship between BPSD and care services use or lifestyle habits. The model was adjusted for sociodemographic factors, and the interaction terms of the pandemic period with lifestyle and service use were included to evaluate the effects of COVID-19. (3) Results: We identified 977 participants with MCI and 1380 with dementia (MCI group: 69.8% age 75 years or older, 54.2% female; dementia group: 79.8% age 75 years or older, 64.8% female). After adjustment for possible confounders, significantly worse BPSD was demonstrated in those who used daycare services during COVID-19 (both MCI and dementia patients; p = 0.007 and p = 0.025 respectively) and in those with poor nutritional function (dementia patients; p = 0.040). (4) Conclusions and Implications: During COVID-19, poor nutritional status and use of daycare services were associated with BPSD in those with cognitive decline. These findings indicate the need to fully examine the quantity and quality of care services for people with cognitive decline during emergencies and to continue to provide effective services.
- 13 June 2022
- Published by Elsevier in Journal of Infection and Chemotherapy
- Vol. 28 (10) , 1439-1444
Risk factors associated with hospitalization in patients with asymptomatic or mild COVID-19 in public accommodation facilities in Tokyo
- 13 June 2022
- Published by Elsevier in Journal of Infection and Chemotherapy
- Vol. 28 (10) , 1439-1444
Risk factors associated with hospitalization in patients with asymptomatic or mild COVID-19 in public accommodation facilities in Tokyo