Fundamentals of machine learning

It is intended that the student understands the main concepts associated with Statistical or Automatic Learning and its mathematical foundations. A tour through the main machine learning techniques and tools that arise from both supervised and unsupervised regression and classification problems will be done, analyzing the advantages, risks and disadvantages of each approach.
The objectives to be achieved as a result of the learning are:
• Identify and model problems in real applications by choosing the appropriate tool.
• Understand the implications of the structural hypotheses in the results of the models and possible reformulations.
• Build statistical learning models for data analysis.
• Interpret the results for presentation to both specialized and non-specialized audiences.

Unit 1. Fundamentals of Probability, Statistics and Statistical Learning. General tools for data preprocessing and model validation.
Unit 2. Classic models of Regression and Classification
Unit 3. Extensions of regression and classification models.
Unit 4. Techniques of resampling and assembly of models.
Unit 5. Unsupervised classification techniques

Basic
• Friedman, J.; Hastie, T. and Tibshirani, R. (2008) The Elements of Statistical Learning. Springer
• James, G.; Witten, D.; Hastie, T. and Tibshirani (2021) An introduction to Statistical Learning 2ed. Springer.

Complementary
• Pathak, M.A. (2014). Beginning Data Science with R. Springer
• Theodoridis, S. (2020). Machine Learning. A Bayesian and Optimization Perspective. Academic Press.

CG2 - Ability to solve problems with initiative, decision-making, autonomy and creativity.
CG4 - Ability to select and justify the appropriate methods and techniques to solve a specific problem, or to develop and propose new methods based on artificial intelligence.
CG5 - Ability to design new computational systems and/or evaluate the performance of existing systems, which integrate artificial intelligence models and techniques.
CB3 – Ability to gather and interpret relevant data (usually within your area of study) to make judgments that include reflection on relevant social, scientific or ethical issues
CB5 – Ability to undertake further studies with a high degree of autonomy
TR3 - Ability to create new models and solutions autonomously and creatively, adapting to new situations. Initiative and entrepreneurial spirit.
CE1 - Ability to use mathematical and statistical concepts and methods to model and solve artificial intelligence problems
CE15 - Know and know how to correctly apply and explain the validation techniques of artificial intelligence solutions.
CE2 - Ability to solve artificial intelligence problems that require algorithms, correctly applying software development and design methodologies.
CE12 - To know the fundamentals of artificial intelligence algorithms and models for solving problems of certain complexity, understand their computational complexity and have the ability to design new models.

The teaching methodology will consist of expositive (20hrs) and interactive (30hrs) classes, as well as tutoring of learning and the tasks assigned to the students. Expositive lectures will consist to expose the contents of the different subjects, with special emphasis on the explanation and assimilation of concepts, mathematical foundations and utility. In the interactive classes, practical problems related to the topics will be solved. The structure of the lectures is designed for covering all the competences.

The final grade will be the combination of the final exam grade with the grade of the continuous assessment in a proportion (60%-40%). The final exam will consist of a computer-based test on solving practical problems. Continuous assessment will consist of tests done in laboratory hours throughout the course. For the second oportunity, the grade of the continuous assessment will be maintained.

Each ECTS credit translates into 8 hours of face-to-face classes. It is estimated that the student will need, for each hour of face-to-face class, an additional hour for the global understanding of the content. The completion of continuous assessment work or the preparation of tests will amount to 9 hours per ECTS credit. A total of 25 hours per ECTS credit will result.

It is important to have fresh concepts of Algebra, Calculus and Numerical Analysis and Statistics from previous courses.

It is advisable to participate actively in the learning process of the subject: attendance and participation in theoretical and computer classes and the use of tutoring hours as well as the realization of a responsible effort in the work and personal assimilation of the methods studied.

For cases of fraudulent completion of exercises and tests, the specific regulation of the University will be applied.
This guide and the criteria and methodologies described in it are subject to modifications arising from USC regulations and guidelines.