Data Analysis and Artificial Intelligence Methods (FSI-TPX)

The student:
• analyzes physical data using statistical and AI methods,
• can apply MLE and hypothesis testing to experimental data,
• uses regression and classification methods (Naive Bayes, SVM, neural networks),
• performs dimension reduction (PCA) and clustering,
• understands the principles of Bayesian and reinforcement learning,
• interprets model results in the context of physical processes

The course develops students' ability to process and analyze physical data using statistical and artificial intelligence methods. Students will learn the principles of parameter estimation, hypothesis testing, classification, regression, and modern machine and reinforcement learning methods.

Syllabus
Week 1: Introduction to Data Analysis and Workflow – Overview of ML and AI in Physics, Pandas and Matplotlib Review, Data Types, Analysis Workflow, Probability
Week 2: Statistical Data Analysis – Mean, Variance, Distribution, Histograms, Uncertainty Estimation
Week 3: Parameter Estimation and Maximum Likelihood (MLE) – Likelihood, Log-Likelihood, Numerical Maximization, Applications to Physics Data
Week 4: Statistical Hypothesis Testing – Principles of Testing, P-Value, Significance, T-Test, χ²-Test, Measurement Agreement Test
Week 5: Regression Methods – Linear and Nonlinear Regression, Least Squares, Relationship to MLE, Quality of Fit
Week 6: Data Classification: Naive Bayes and Support Vector Machines (SVM) – Supervised Learning, Kernel Functions, Applications in Physics
Week 7: Clustering and PCA Methods – Unsupervised Learning, k-means, PCA – dimension reduction, visualization of physical data
Week 8: Neural networks – perceptron architecture, activation functions, training, TensorFlow/Keras
Week 9: Regression and classification using neural networks – prediction of physical quantities, regularization, overfitting
Week 10: Bayesian modeling – Bayesian inference, MAP estimation, relation to MLE, interpretation in physics
Week 11: Reinforcement Learning (basics) – agent, environment, reward, simple RL simulations (e.g. pendulum stabilization)
Week 12: Project: Analysis of physical data using ML/AI – individual or team work, presentation and interpretation of results

VanderPlas, J. Python Data Science Handbook. Sebastopol: O’Reilly Media, 2016. ISBN 978-1491912058.

Géron, A. Hands-On Machine Learning with Scikit-Learn and TensorFlow. Sebastopol: O’Reilly Media, 2017. ISBN 978-1491962299.

Barber, D. Bayesian Reasoning and Machine Learning. Cambridge: Cambridge University Press, 2012. ISBN 978-0521518147.

Scikit-learn documentation [online]. The scikit-learn developers, 2010–2025. Dostupné z: https://scikit-learn.org/

TensorFlow documentation [online]. Google, 2015–2025. Dostupné z: https://www.tensorflow.org/

scipy.stats documentation [online]. SciPy Developers, 2001–2025. Dostupné z: https://docs.scipy.org/

CERN Open Data Portal [online]. CERN, 2014–2025. Dostupné z: https://opendata.cern.ch/

NASA Open Data Portal [online]. NASA, 2011–2025. Dostupné z: https://data.nasa.gov/