Мы используем файлы cookies для улучшения работы сайта НИУ ВШЭ и большего удобства его использования. Более подробную информацию об использовании файлов cookies можно найти здесь, наши правила обработки персональных данных – здесь. Продолжая пользоваться сайтом, вы подтверждаете, что были проинформированы об использовании файлов cookies сайтом НИУ ВШЭ и согласны с нашими правилами обработки персональных данных. Вы можете отключить файлы cookies в настройках Вашего браузера.

  • A
  • A
  • A
  • АБВ
  • АБВ
  • АБВ
  • А
  • А
  • А
  • А
  • А
Обычная версия сайта
14
Сентябрь

Introduction to multiple hypothesis testing procedures

2025/2026
Учебный год
ENG
Обучение ведется на английском языке

Преподаватель

Course Syllabus

Abstract

This course examines multiple hypothesis testing procedures that allow you to answer several interrelated questions based on available observations or data, while monitoring some indicators of quality and the procedure used.
Learning Objectives

Learning Objectives

  • Recall the classical results of testing statistical hypotheses.Memorize the basic concepts of the theory of constructing multiple hypothesis testing procedures; apply the test obtained in the lecture to solve the problem on the basis of dispersion analysis; memorize the method of unification and intersection; study the concept of a hierarchical family; apply the LSD procedure in practice; study the closure principle of constructing multiple hypothesis testing procedures; apply the Hill procedure in practice; study the principle of partitioning for constructing procedures that control the probability of at least one false statement (FWER). Examine procedures that control the proportion of false claims (FDR). To apply the Hochberg procedure in practice; to study the procedures for choosing one of many hypotheses; to study what errors of the 1st and 2nd kind and risk and loss functions are; to count in practice the number of errors of the 1st and 2nd kind and the risk of the procedure; to study the Bayesian approach to multiple hypothesis testing. To put into practice the Bayesian approach to multiple hypothesis testing.
Expected Learning Outcomes

Expected Learning Outcomes

  • To apply classical approaches to hypothesis testing against alternatives in practice.
  • Recall classical approaches to testing hypotheses against alternatives.
  • To put into practice the test obtained in the lecture to solve the problem of variance analysis
  • Memorize the basic concepts of the theory of constructing multiple hypothesis testing procedures
  • Put the LSD procedure into practice
  • To study the concept of a hierarchical family
  • Memorize the method of union and intersection
  • Put the Hill procedure into practice.
  • To study the principle of closure of the construction of multiple hypothesis testing procedures.
  • To put into practice the maximum procedure of multiple hypothesis testing.
  • To study the optimal multiple hypothesis testing procedure controlling FWER.
  • To put the Hochberg procedure into practice.
  • To study the principle of partitioning to build procedures that control the probability of at least one false statement (FWER).
  • Examine procedures that control the proportion of false claims (FDR).
  • Calculate in practice the number of errors of the 1st and 2nd kind and the risk of the procedure.
  • To study what errors of the 1st and 2nd kind and functions of risk and loss are.
  • To study the procedures for choosing one of the many hypotheses.
  • To put Bayesian approach to multiple hypothesis testing into practice.
  • To study the Bayesian approach to multiple hypothesis testing.
Course Contents

Course Contents

  • The principle of closure.
  • Optimal maximal procedure for multiple hypothesis testing.
  • The union-intersection method. A hierarchical family.
  • An introduction to multiple hypothesis testing.
  • The two-solution rule. The Neiman-Pearson approach.
  • The principle of partitioning for constructing procedures that control the probability of at least one false statement (FWER). Procedures that control the proportion of false claims (FDR).
  • Procedures for selecting one of many hypotheses. The function of risk and loss. The Wald-Lehman theory.
  • The Bayesian approach to multiple hypothesis testing.
Assessment Elements

Assessment Elements

  • non-blocking Control work 1
  • non-blocking Control work 2
  • non-blocking Control work 3
  • non-blocking Exam
Interim Assessment

Interim Assessment

  • 2025/2026 2nd module
    0.5* mark for practice+ 0.5* mark for theory
Bibliography

Bibliography

Recommended Core Bibliography

  • Basics of modern mathematical statistics : exercises and solutions, , 2014
  • Graphical models with R, Hojsgaard, S., 2012
  • Introduction to mathematical statistics and its applications, Larsen, R. J., 2014
  • Introduction to mathematical statistics, Hogg, R. V., 2014
  • Mathematical statistics with applications in R, Ramachandran, K. M., 2015
  • Mathematical statistics with applications, Wackerly, D. D., 2008
  • Probability theory and statistical inference : econometric modeling with observational data, Spanos, A., 2000
  • Probability theory and statistical inference : econometric modeling with observational data, Spanos, A., 2004
  • Теория вероятностей и математическая статистика, Колданов, А. П., 2023
  • Теория вероятностей и математическая статистика, учебник, 243 с., Колданов, А. П., Колданов, П. А., 2023

Recommended Additional Bibliography

  • Advances in inequalities from probability theory and statistics, Barnett, N. S., 2008

Authors

  • Koldanov Petr Aleksandrovich