Teoretyczne Podstawy Informatyki

Advanced Methods for Data Analysis

Wydział Fizyki, Astronomii i Informatyki Stosowanej,

Uniwersytet Jagielloński w Krakowie

Rok akademicki 2024/2025

Konsultacje: czwartek, godz. 10:00 - 11:00; pokój G-0-10.

Exams:
--------------------

     29.01.2025, A-1-03, 11.00-13.00
     25.02.2025, A-1-03,   9.00-11.00

Lectures
---------
Recommended books/articles for reading:

=> G. Cowan, "Statistical Data Analysis"
      recorded lectures at CERN Summer School 2023: link
=> F. James, "Statistical Methods in Experimental Physics"
=> J. Narsky, F. Porter, "Statistical Analysis Techniques in Particle Physics"
=> J. A.Rice, "Mathematical Statistics and Data Analysis"
=> I. Narsky, F. C. Porter, "Statistical Analysis Techniques in Particle Physics"
=> K. Cranmer, "Practical statistics for LHC"

Recent conferences and workshops:

=> PHYSTAT (2024): Statistics Meets Machine Learning
=> PHYSTAT (2024): Unfolding
=> PHYSTAT (2024): Simulation Based Inference in Fundamental Physics
=> PHYSTAT (2022): Anomalies

Date	Lecture slides	Additional material
	Statistics and Data Analysis: Basic	Statistical methods for LHC (advanced)
10.10.2024	Introduction, StatAnal-lecture-1	Efficiency uncertainties
17.10.2024	StatAnal-lecture-2
24.10.2024	StatAnal-lecture-3	EW measurements at LEP2
31.10.2024	Canceled (Rector's hours)
7.11.2024	StatAnal-lecture-4
	Statistics and Data Analysis: Advanced
14.11.2024	LHCStatAnal-lecture-1	https://arxiv.org/pdf/1007.1727.pdf https://arxiv.org/pdf/1609.04150.pdf https://arxiv.org/pdf/1807.05996.pdf https://arxiv.org/pdf/2101.06944.pdf HistFactory Pyhf https://arxiv.org/pdf/2109.04981.pdf
21.11.2024	LHCStatAnal-lecture-2 LHCStatAnal-lecture-3
	Statistics and Data Analysis: Questions for exam
	Multivariate Techniques and Machine Learning
28.11.2024	Unfolding-lecture	https://arxiv.org/pdf/1910.14654.pdf F. Spano-Proc14-02/P52.pdf https://arxiv.org/pdf/1611.01927.pdf
5.12.2024	MVandML-lecture-1	https://arxiv.org/pdf/1506.02169.pdf P.Bhat, Multivariate_Analysis_Methods_in_Particle_Physics Understanding Deep Learning https://arxiv.org/pdf/1806.11484.pdf ML4Jets 2024 workshop https://atlas.cern/Updates/Feature/Machine-Learning
12.12.2024	MVandML-lecture-1a, MVandML-lecture-1b, MVandML-lecture-1c postponed and moved to on-line starts 9.01.2025 at 18:00 pm	https://iopscience.iop.org/article/ 10.1088/1748-0221/11/01/P01019/pdf https://arxiv.org/pdf/1812.09722.pdf ATL-PHYS-PUB-2019-033.pdf ATL-PHYS-PUB-2020-018.pdf GNN in ATLAS flavour tagging
19.12.2024	MVandML-lecture-2
	MV and ML: Questions for exam
	Physics Modeling, Simulation and Monte Carlo Methods
9.01.2025	PhysModel-lecture	https://www.coursera.org/learn/ modeling-simulation-natural-processes
16.01.2025	MCandSimulation-lecture
23.01.2025	MCandSimulation-lecture
	Phys. Model and MC methods: Questions for exam (??)

Assignments:
-------------------------

Date	Topic	Root/C++ or use PyRoot	Datasets/Tutorials	Python + Anaconda	Datasets/Tutorials
	Statistics and Data Analysis
10.10.2024	Introduction-labs StatAnal_labs-lecture-1.txt				Scripts in K. Cranmer, Statistics and Data Science
17.10.2024	StatAnal_labs-lecture-2.txt
24.10.2024	StatAnal_labs-lecture-3.txt
31.10.2024	Canceled (Rector's hours)
7.11.2024	StatAnal_labs-lecture-4.txt
14.11-5.12.2024	StatAnal-project: select one from suggested topics or propose your own.	1) Modeling tools: RooFit, RooStats and HistFactory LHCStatAnal-labs-4-Root follow exercises there 2) Interval Estimation and Hypotheses Testing by T. Dorigo, IN2P3 School on Statistics, 2018 slides follow exercises there 3) Higgs signal at LHC by I. van Vulpen, Terascale Statistics School, DESY 2018 slides and exercises DesyCode2018.tgz follow exercises there		1) Folow examples in: Asia-Europe Pacific School 2022 lectures+hands-on by N. Berger Practical Statistics-1 Practical-Statistics-2 Practical-Statistics-3 2) Try out PYHF tool on exercises proposed for Rootfit/Roostats PYHF: python based fitting/limit-setting/interval estimation
	Multivariate techniques and Machine Learning
12.12.2024 - 9.01.2025	MVandML-project: select one from suggested topics or propose your own.	1) BDTs and TMVA by I. Coadou, IN2P3 School on Statistics, 2018 slides Apply.C Train.C dataSchachbrett.root follow exercises there 2) Unfolding RooUnfold slides follow exercises there 3) Analysis of ATLAS open data with MV or ML methods 4) Analysis with MVA methods by N. Chanon (ETH Zurich, 2012)		1) Analysis of ATLAS open data BDT example for H->4l infofile.py 2)Unfolding with Gaussian processes: https://arxiv.org/pdf/1811.01242.pdf https://github.com/adambozson/gp-unfold 3) Follow hands-on examples/course https://jduarte.physics.ucsd.edu/phys139_239/README.html
	Physics Modeling, Simulation and Monte Carlo Methods
16.01.2025 -30.01.2025	PhysModelandMC-project:

Aachen Online Statistics School 2023
https://indico.desy.de/event/37562/timetable/

Statistical Analysis in HEP Physics
N. Beger, Foundation of Statistics, Lectures at CERN Summer School 2019
link1, link2, link3

Statistics and Data Science
K. Cranmer, Course at NYU Physics, Fall 2020, link

Machine learning applications in HEP physics:

B. Nachman,
"Advanced Machine Learning for Classification, Regression, and Generation in Jet Physics"

M. Stoye,
"ML applications in CMS"

ML techniques in HEP, Workshop, Berkeley Laboratory, 11 - 13 December 2018
https://indico.physics.lbl.gov/indico/event/546/

A. Castaneda, LHCP conference, Puebla, Mexico, 2019
ML and Big data tools at HEP,

Last part of the course will be based on the materials from Coursera:

Dr. Mine Çetinkaya-Rundel "Data Analysis and Statistical Inference"
C. Guestrin and E. Fox, "Machine Learning Specialisation"
        Foundation: link
        Regression: link
        Classification: link
        Clustering and Retrieval: link

Related interesting material from Coursera
D. Peng, J. Leek and B. Caffo, " Exploratory Data Analysis"
J. Leskovec, A. Rajaraman and J. Ullman, "Mining Massive Datasets"
B. Caffo, R. D. Peng and J. Leek, "Regression Models"
B. Chopard et al., "Simulation and modeling of natural processes"

Collection of datasets
http://mlr.cs.umass.edu/ml/datasets.html
http://faculty.marshall.usc.edu/gareth-james/ISL/data.html
http://snap.stanford.edu/data/amazon/

Useful links:
https://turi.com/download/install-graphlab-create-aws-coursera.html
https://turi.com/download/academic.html
https://github.com/turi-code/SFrame

Clustering
http://scikit-learn.org/stable/modules/generated/sklearn.feature_extraction.text.CountVectorizer.html
http://scikit-learn.org/stable/modules/generated/sklearn.feature_extraction.text.TfidfVectorizer.html

Boosting
https://turi.com/learn/userguide/supervised-learning/boosted_trees_classifier.html
https://homes.cs.washington.edu/~tqchen/pdf/BoostedTree.pdf
http://scikit-learn.org/stable/modules/generated/sklearn.ensemble.GradientBoostingClassifier.html
-----------------
[1] https://class.coursera.org/statistics
[2] http://www.openintro.org/stat/textbook.php
[3] https://class.coursera.org/exdata-006
[4] https://class.coursera.org/mmds
[5] http://www.mmds.org/
[6] http://www.cs.cmu.edu/~awm/tutorials.html

Additional materials
http://www.stat.cmu.edu/~cshalizi/ADAfaEPoV/ADAfaEPoV.pdf
http://statweb.stanford.edu/~tibs/ElemStatLearn/
http://en.wikipedia.org/wiki/Determining_the_number_of_clusters_in_a_data_set
http://www.youtube.com/watch?v=wQhVWUcXM0A

Ostatnia modyfikacja: 7 October 2024

Elzbieta Richter-Was

Wstecz

PYHF: python based fitting/limit-setting/interval estimation