Machine Learning Bio
The following are last-minute news you should be aware of ;-)
* 23/02/2022: Lectures start today!
Course Aim & Organization
The objective of the Machine Learning course is to give an in-depth presentation of the techniques most used for pattern recognition, knowledge discovery, and data analysis/modeling. These techniques are presented both from a theoretical (i.e., statistics and information theory) perspective and a practical one (i.e., coding examples) through the descriptions of algorithms and their implementations in a general-purpose programming language (i.e., python).
The course presents the classical supervised and unsupervised learning paradigms described and discussed presenting regression, classification, and clustering problems in Bioinformatics. The course is composed of a set of lectures on specific machine learning techniques (e.g., generalized linear regression, logistic regression, linear and quadratic discriminant analysis, support vector machines, k-nearest-neighborhood, clustering, etc.) preceded by the introduction of the Statistical Learning framework which acts as a common reference framework for the entire course.
The course is composed of a blending of lectures and exercises by the course teacher and a teaching assistant.
- Matteo Matteucci: the course teacher and here it is his webex room
- Marco Cannici: the teaching assistant and here it is his webex room
The course mostly follows the following book which is also available for download in pdf
- An Introduction to Statistical Learning with Applications in R by Gareth James, Daniela Witten, Trevor Hastie, and Robert Tibshirani
The course lectures will present the theory and practice of the following:
- Machine Learning and Pattern Classification: the general concepts of Machine Learning and Pattern Recognition are introduced within the framework of Statistical Decision Theory with reference to the bias-variance trade-off and the Bayes classifier;
- Generalized Linear Regression: linear methods for regression will be presented and discussed introducing different techniques (e.g., Linear Regression, Ridge Regression, K-Nearest Neighbors Regression, Non-Linear Regression, etc.) and the most common methodologies for model validation and selection (e.g., AIC, BIC, cross-validation, stepwise feature selection, Lasso, etc.).
- Linear and Non-Linear Classification: generative and discriminative techniques for classification will be described and discussed (e.g., Logistic Regression, Linear and Quadratic Discriminant Analysis, K-Nearest Neighbors, Perceptron Rule, and Support Vector Machines, etc.). Metrics for classifiers evaluation and comparison are presented in this part of the course (e.g., accuracy, precision, recall, ROC, AUC, F-measure, Matthew coefficient).
- Unsupervised Learning: the most common approaches to unsupervised learning are described mostly focusing on clustering methods such as hierarchical clustering, k-means, k-medoids, Mixture of Gaussians, DBSCAN, etc
These topics will be presented both from a theoretical perspective and a practical one via implementations in the general-purpose programming language python.
Detailed course schedule
A detailed schedule of the course can be found here; topics are just indicative while days and teachers are correct up to some last-minute change (I will notify you by email). Please note that not all days we have lectures!!
Chapters are intended as complete except for
- Ch.4 ESL: Section 4.5
- Ch.12 ESL: Sections 12.1, 12.2, 12.3
- Ch.9 ISL: Sections 9.1, 9.2, 9.3
The course evaluation is composed by two parts:
- HW: Homework with exercises covering the whole program (up to 6 points)
- OR: An oral discussion covering the whole program (up to 26 points)
the final score will be the sum of HW (not compulsory) and OR scores. You will get the oral grade as a mark in the scale of 30 up to 32/30 which you have to multiply by 0.8125 and then you add to it the score of the project. For your convenience here it is a conversion table with the final mark in case you do not turn in the project or you get the whole 6 marks in the project.
Teaching Material (the textbook)
Lectures will be based on material taken from the book.
- An Introduction to Statistical Learning with Applications in R by Gareth James, Daniela Witten, Trevor Hastie and Robert Tibshirani
If you are interested in a more deep treatment of the topics you can refer to the following book from the same authors
- The Elements of Statistical Learning: Data Mining, Inference, and Prediction. by Trevor Hastie, Robert Tibshirani, and Jerome Friedman.
Some additional material that could be used to prepare the oral examination will be provided together with the past homeworks.
In the following you can find the lecture slides used by the teacher and the teaching assistants during classes.
- [2020/2021] Linear Algebra Basics: Basic elements of linear algebra, e.g., vectors, matrices, basis, span, etc.
- [2020/2021] Course introduction: introductory slides of the course with useful information about the grading, and the course logistics. Some examples from supervised and unsupervised learning. Regression, classification, clustering terminology and examples.
- [2020/2021] Statistical Learning Introduction: Statistical Learning definition, rationale, and trade-offs (e.g., prediction vs. inference, parametric vs non parametric models, flexibility vs. interpretability, etc.)
- [2020/2021] Linear Regression: Simple Linear Regression and Multiple Linear Regression. Generalized Linear models. Cross-validation techniques. Feature selection. Ridge Regression and Lasso.
- [2020/2021] Linear Classification: From Linear Regression to Logistic Regression. Linear Discriminant Analysis and Quadratic Discriminant Analysis. Comparison between linear classification methods. Discriminative vs. generative methods. Support Vector Machines.
- [2020/2021] Clustering: Introduction to unsupervised learning and clustering, hierarchical clustering, k-means, DBSCNA, indexes for clustering evaluation.
- [2020/2021] Principal Component Analysis: Principal Component Analysis, Geometric Interpretation, Singular Values Decomposition.
We will use Python (with Jupyter notebooks) throughout the course, thus we kindly ask you to install the "Anaconda" package to be ready for the labs. Here are the download links. You can find a simple Jupyter notebook HERE to test if the installation succeeded.
To open the notebook
- launch the "Anaconda Navigator" app
- launch the "jupyter Notebook" app within the navigator, it should automatically open a webpage (it may take a while)
- on the webpage, navigate on the folder where you downloaded the "lab01.00-TestEnvironment.ipynb" file, and press on the file to open it
- then follow the instruction within the notebook
If you didn't install anaconda but just jupyter, or if you can't find Anaconda Navigator
- open a shell ("Anaconda Prompt" if you are using Windows)
- move on to the folder where you downloaded "lab01.00-TestEnvironment.ipynb"
- run the "jupyter notebook" command, it should automatically open a webpage
- then on the webpage press on the "lab01.00-TestEnvironment.ipynb" file and follow the instructions within the notebook
The following are the notebook used in the labs:
Papers and links useful to integrate the textbook
- Basic Linear Algebra: "Basic Linear Algebra" chapter from Wayne Winston book "Operations Research Applications and Algorithms (4th ed.)"
- Bias vs. Variance: "Understanding the Bias-Variance Tradeoff" essay by Scott Fortmann-Roe
- Karush Kuhn Tucker Conditions: a short note on their meaning with references to relevant wikipedia pages
- Seeing Theory: a website where the basic concepts of probability and statistics are explained in a visual way.
Python examples to better practice with numpy library
- Some numpy exercises, we did roughly 1, 2, 4, 6, 9, 10, 13
- Othen numpy exercises from simple to complex
- More numpy exercises from simple to complex
The following are links to online sources which might be useful to complement the material above
- An Introduction to Linear Algebra with numpy examples. It provides the very fundamental definitions, does not cover eigenvalues and eigenvectors.
- Statistical Learning MOOC covering the entire ISL book offered by Trevor Hastie and Rob Tibshirani. Start anytime in self-paced mode.
- MATH 574M University of Arizona Course on Statistical Machine Learning and Data Mining; here you can find slides covering part of the course topics (the reference book for this course is again The Elements of Statistical Learning)