2024
Machine Learning
Name: Machine Learning
Code: INF13203L
6 ECTS
Duration: 15 weeks/156 hours
Scientific Area:
Informatics
Teaching languages: Portuguese
Languages of tutoring support: Portuguese
Regime de Frequência: Presencial
Sustainable Development Goals
Learning Goals
At the end of the course unit the student will:
* understand the fundamental concepts of machine learning
* know a wide range of machine learning approaches and algorithms, namely supervised learning
* understand the various stages in the construction of an intelligent system and what are the techniques that can be applied in each step (definition of the problem, feature extraction, creation of training, test and validation sets, algorithm application, performance evaluation)
* know how to design/program a machine learning system
* understand the fundamental concepts of machine learning
* know a wide range of machine learning approaches and algorithms, namely supervised learning
* understand the various stages in the construction of an intelligent system and what are the techniques that can be applied in each step (definition of the problem, feature extraction, creation of training, test and validation sets, algorithm application, performance evaluation)
* know how to design/program a machine learning system
Contents
Basic concepts
Machine Learning paradigms: supervised, unsupervised, re-inforcement learning
Supervised learning: classification and regression
Binary, multi-class and multi-label classification
Algorithms: logistic regression, perceptron, decision trees, rules, naive Bayes, support vector machines
ML practice: overfitting, bias/variance tradeoff, model selection (train/test, holdout, cross-validation), confusion matrix and evaluation metrics (accuracy, error, precision, recall, others)
Unsupervised learning: clustering
Algorithms: K-means, EM
Clustering evaluation measures
Introduction to ensemble methods
Machine Learning paradigms: supervised, unsupervised, re-inforcement learning
Supervised learning: classification and regression
Binary, multi-class and multi-label classification
Algorithms: logistic regression, perceptron, decision trees, rules, naive Bayes, support vector machines
ML practice: overfitting, bias/variance tradeoff, model selection (train/test, holdout, cross-validation), confusion matrix and evaluation metrics (accuracy, error, precision, recall, others)
Unsupervised learning: clustering
Algorithms: K-means, EM
Clustering evaluation measures
Introduction to ensemble methods
Teaching Methods
Teaching methodologies:
Theoretical classes; lab classes with problems that accompany the theoretical material.
Availability of exercises, of gradual difficulty, covering the topics taught, for students to practice mastery of the subject.
Evaluation:
(i) two written frequencies and / or (ii) final written exam
(iii) individual and group exercises and (iv) development of a small project
Theoretical classes; lab classes with problems that accompany the theoretical material.
Availability of exercises, of gradual difficulty, covering the topics taught, for students to practice mastery of the subject.
Evaluation:
(i) two written frequencies and / or (ii) final written exam
(iii) individual and group exercises and (iv) development of a small project
Assessment
Continuous assessment:
theory: 50% - (i) two written frequencies (25% each)
labs: 50% - (iii) implementation assignment (25%) and (iv) ML challenge (25%)
Final assessment:
theory: 50% - (ii) final written exam (50%)
labs: 50% - (iii) implementation assignment (25%) and (iv) ML challenge (25%)
theory: 50% - (i) two written frequencies (25% each)
labs: 50% - (iii) implementation assignment (25%) and (iv) ML challenge (25%)
Final assessment:
theory: 50% - (ii) final written exam (50%)
labs: 50% - (iii) implementation assignment (25%) and (iv) ML challenge (25%)
Teaching Staff
- Teresa Cristina de Freitas Gonçalves [responsible]
Certificações
