Hayashi, Kenichi

写真a

Affiliation

Faculty of Science and Technology, Department of Mathematics (Yagami)

Position

Associate Professor
Page Top ▲
 

Research Areas 【 Display / hide 】

  • Informatics / Statistical science

Page Top ▲
 

Books 【 Display / hide 】

  • Rで学ぶ統計的データ解析

    林 賢一, 講談社サイエンティフィク, 2020.11,  Page： 352

  • Machine Learning: The Art and Science of Algorithms that Make Sense of Data

    竹村彰通（監訳），田中研太郎，小林景，兵頭昌，片山翔太，山本倫生，吉田拓真，林賢一，松井秀俊，小泉和之，永井勇, 朝倉書店, 2017.04

    Scope: 7章

Page Top ▲

Papers 【 Display / hide 】

  • Asymptotic Properties of Matthews Correlation Coefficient

    Itaya Y., Tamura J., Hayashi K., Yamamoto K.

    Statistics in Medicine 44 ( 1-2 )  2025.01

    ISSN  02776715

    　View Summary

    Evaluating classifications is crucial in statistics and machine learning, as it influences decision-making across various fields, such as patient prognosis and therapy in critical conditions. The Matthews correlation coefficient (MCC), also known as the phi coefficient, is recognized as a performance metric with high reliability, offering a balanced measurement even in the presence of class imbalances. Despite its importance, there remains a notable lack of comprehensive research on the statistical inference of MCC. This deficiency often leads to studies merely validating and comparing MCC point estimates—a practice that, while common, overlooks the statistical significance and reliability of results. Addressing this research gap, our paper introduces and evaluates several methods to construct asymptotic confidence intervals for the single MCC and the differences between MCCs in paired designs. Through simulations across various scenarios, we evaluate the finite-sample behavior of these methods and compare their performances. Furthermore, through real data analysis, we illustrate the potential utility of our findings in comparing binary classifiers, highlighting the possible contributions of our research in this field.

  • Optimal machine learning models for developing prognostic predictions in patients with advanced cancer

    Hamano, J., Takeuchi, A., Keyaki, T., Nose, H., Hayashi, K.

    Cereus 16 ( 12 )  2024.12

    Research paper (scientific journal), Joint Work, Last author, Accepted

  • A new integrated discrimination improvement index via odds

    Hayashi K., Eguchi S.

    Statistical Papers 65 ( 8 ) 4971 - 4990 2024.10

    ISSN  09325026

    　View Summary

    Consider adding new covariates to an established binary regression model to improve prediction performance. Although difference in the area under the ROC curve (delta AUC) is typically used to evaluate the degree of improvement in such situations, its power is not high due to being a rank-based statistic. As an alternative to delta AUC, integrated discrimination improvement (IDI) has been proposed by Pencina et al. (2008). However, several papers have pointed out that IDI erroneously detects meaningless improvement. In the present study, we propose a novel index for prediction improvement having Fisher consistency, implying that it overcomes the problems in both delta AUC and IDI. Furthermore, our proposed index also has an advantage that the index we proposed in our previous study (Hayashi and Eguchi 2019) lacked: it does not require any hyperparameters or complicated transformations that would make interpretation difficult.

  • Model Selection with Missing Data Embedded in Missing-at-Random Data

    Takai K., Hayashi K.

    Stats (Stats)  6 ( 2 ) 495 - 505 2023.06

    Joint Work, Accepted

    　View Summary

    When models are built with missing data, an information criterion is needed to select the best model among the various candidates. Using a conventional information criterion for missing data may lead to the selection of the wrong model when data are not missing at random. Conventional information criteria implicitly assume that any subset of missing-at-random data is also missing at random, and thus the maximum likelihood estimator is assumed to be consistent; that is, it is assumed that the estimator will converge to the true value. However, this assumption may not be practical. In this paper, we develop an information criterion that works even for not-missing-at-random data, so long as the largest missing data set is missing at random. Simulations are performed to show the superiority of the proposed information criterion over conventional criteria.

  • An Accelerated Failure Time Cure Model with Shifted Gamma Frailty and Its Application to Epidemiological Research

    Aida H., Hayashi K., Takeuchi A., Sugiyama D., Okamura T.

    Healthcare (Switzerland) (Healthcare (Switzerland))  10 ( 8 )  2022.08

    Accepted

    　View Summary

    Survival analysis is a set of methods for statistical inference concerning the time until the occurrence of an event. One of the main objectives of survival analysis is to evaluate the effects of different covariates on event time. Although the proportional hazards model is widely used in survival analysis, it assumes that the ratio of the hazard functions is constant over time. This assumption is likely to be violated in practice, leading to erroneous inferences and inappropriate conclusions. The accelerated failure time model is an alternative to the proportional hazards model that does not require such a strong assumption. Moreover, it is sometimes plausible to consider the existence of cured patients or long-term survivors. The survival regression models in such contexts are referred to as cure models. In this study, we consider the accelerated failure time cure model with frailty for uncured patients. Frailty is a latent random variable representing patients’ characteristics that cannot be described by observed covariates. This enables us to flexibly account for individual heterogeneities. Our proposed model assumes a shifted gamma distribution for frailty to represent uncured patients’ heterogeneities. We construct an estimation algorithm for the proposed model, and evaluate its performance via numerical simulations. Furthermore, as an application of the proposed model, we use a real dataset, Specific Health Checkups, concerning the onset of hypertension. Results from a model comparison suggest that the proposed model is superior to existing alternatives.

display all >>

Page Top ▲

Papers, etc., Registered in KOARA 【 Display / hide 】

Page Top ▲

Reviews, Commentaries, etc. 【 Display / hide 】

Page Top ▲

Research Projects of Competitive Funds, etc. 【 Display / hide 】

  • 不均衡データに対する解析法の統合的理解と生存時間解析への発展的応用

    2023.04
    -
    2027.03

    基盤研究(C), Principal investigator

  • 異質な集団を含むデータに対する統計的学習理論を用いたモデル開発と臨床医学への応用

    2018.04
    -
    2022.03

    MEXT,JSPS, Grant-in-Aid for Scientific Research, Grant-in-Aid for Scientific Research (C), Principal investigator

  • A study on exploratory identification of responsive subgroups in complex survival event data

    2015.04
    -
    2018.03

    MEXT,JSPS, Grant-in-Aid for Scientific Research, Grant-in-Aid for Young Scientists (B), Principal investigator

Page Top ▲

Awards 【 Display / hide 】

  • 最優秀賞（ポスター部門）

    岡田和也，林賢一, 2024.01,  2023年度スポーツデータサイエンスコンペティション事務局, 2023年度スポーツデータサイエンスコンペティション

    Type of Award： Award from Japanese society, conference, symposium, etc.

    　View Description

    岡田和也（理工学部数理科学科，筆頭著者）との共同研究

  • 若手優秀発表賞

    2021.05, 日本計量生物学会

    Type of Award： Award from Japanese society, conference, symposium, etc.

    　View Description

    2021年度計量生物学会年会 若手優秀発表賞（正会員の部）

  • 最優秀賞

    奥富航，林賢一, 2018.01, 第7回スポーツデータ解析コンペティション審査会, 第7回スポーツデータ解析コンペティション

    Type of Award： Award from Japanese society, conference, symposium, etc.

    　View Description

    奥富航（理工学部数理科学科，筆頭著者）との共同研究

  • 論文賞

    2017.12, 日本分類学会, 日本分類学会

    Type of Award： Award from Japanese society, conference, symposium, etc.

  • 論文賞

    2015.05, 日本計算機統計学会, 日本計算機統計学会

    Type of Award： Award from Japanese society, conference, symposium, etc.

Page Top ▲
 

Courses Taught 【 Display / hide 】

  • TOPICS IN STATISTICAL SCIENCES C

    2025

  • STATISTICAL SCIENCE AND ITS EXERCISE

    2025

  • SEMINAR IN STATISTICAL SCIENCES

    2025

  • MATHEMATICS 2B

    2025

  • MATHEMATICS 1A

    2025

display all >>

Page Top ▲