Welcome to the statistics course page

PRE1 Applied statistics, Sept.-Oct. 2020

AI Master - Université Paris Saclay


Students who analyze data, or aspire to develop new methods for analyzing data should be well trained in statistical methodology. This class aims at teaching the fundamentals of statistical methodology : estimation, confidence intervals and hypothesis testing, including methods based on likelihood and boostrapping. We concentrate more on the understanding of the methods than on the mathematical details. The class include practical exercises, simulation experiments and data analyses with python.

The lecture is given jointly by Marie-Anne Poursat and Timothée Mathieu, Laboratoire de Mathématiques d'Orsay and Celeste team INRIA Sacla
  • marie-anne.poursat@universite-paris-saclay.fr
  • timothee.mathieu@universite-paris-saclay.fr

Thursday 22th of october:
9h-10h TEST3: mandatory to obtain your final grade
10h30-12h Practical session on hypothesis testing.

Information about Test3: this test will last 1 hour, it will contain five methodological questions and a practical exercise.

Training exercises for Test3

Final grading : 30% Test 3  + 70% mean of the 3 best marks among {test1, test2, Lab1, Lab2}

Don't forget to post questions/comments on the forum:
Forum link : https://appliedstatistics.flarum.cloud/
                       Id: Student Pass: ParisSaclay


Course material
Complements on QQplots, ECDF, Bootstrap, Summarizing data: Section 8.5.3 and Chapter 10 of Rice's book

Course 1 :  slidesnotebook
  • homework: read Wasserman's book section 2.1 to 2.4, 2.5 (ex and def 2.18 to 2.20) and 2.7 (ex and def 2.29 to 2.32), 3.1 to 3.3;know how to do questions 1 to 7 of the notebook corrected notebook
Course 2 : slides, notebook, Training exercises
Course 3 : handwritten lecture notes notebook
Course 4 :handwritten lecture notes, notebook
Course 5 : Introduction to the bootstrap (section 8.1-3 of Wasserman's book)
                
handwritten lecture notes,  notebook
Course 6: Hypothesis testing: read sections 10.1-2 of Wasserman's book

Grading scheme

3 time-limited tests (1 written exercise or/and 1 lab exercise) on weeks 3 (sept 24), 5 (oct 8) and 7 (oct 22) : 60% of the final grade
2 small projects (labwork) : 40% of the final grade; the first one given on Sept 25 and due on Oct 2 before 19h, the second one given on Oct 10 and due on Oct 17 before 19h.

Tentative program
  • Course 1: Review of basics in probability and statistics : random variables, probability distributions, descriptive statistics
  • Course 2: Modeling data and fitting probability distributions
  • Course 3: Statistical inference: maximum likelihood estimation
  • Course 4: Laws of ML estimators: variance estimation, approximate distributions, confidence intervals
  • Course 5: Bootstrap
  • Course 6 and 7:Testing hypothesis

References
  • All of statistics: A Concise Course in Statistical Inference, Larry Wasserman
  • Think Stats, Exploratory Data Analysis in Python, Allen B. Downey
  • Mathematical statistics and data analysis, John A. Rice