Análise de Dados

Basic knowledge of intelligent data analysis.

Basic knowledge of Python.

The final grade will be calculated based on the following 3 evaluation components:

1. Assessment in Practical Classes: throughout the semester, challenges will be launched during practical classes related to the subjects being taught. Challenges/exercises will be completed and submitted during class. There will be 6 individual challenges, and the 4 best grades will be considered. This component accounts for 15% of the final grade.
2. Practical assignment/Seminar: the assignment consists of studying, understanding and presentation of a deep learning model. In this work it is intended that students investigate and analise state-of-the-art solutions to computational learning problems. As part of the work, a report will be produced with the main conclusions and a presentation will be made in class. The wok assignment will be disclosed in practical class 10 and the delivery and presentation will take place in the last week of classes. The work has a weight of 35% in the final grade and is carried out in groups of 2 students. The work can be delivered only once, according to the date indicated in the assignment. The grade obtained is valid for all seasons, including the special season. This component has no minimums and is not subject to grade improvement.
3. Written exam: this exam focuses on the theoretical and theoretical-practical component of the course. The exam has a minimum of 35% of the test score.

There are 2 possibilities for completing evaluation:

• Continuous and Periodic Assessment: Students perform the 3 assessment components. In this case, the exam has a weight of 50% in the final grade.
• Periodic Assessment: Students perform only components B and C (practical assignment and exam). In this case, the exam has a weight of 65% in the final grade.

The main objective of the Data Analysis course is to present a set of techniques and methodologies based on deep neural networks to solve real data analysis problems. Understanding the potential of these systems and the ability to develop architectures with these characteristics will give students a set of specialized skills to work in the area of computational data analysis. The main learning objectives of this course are:

1. Know and understand advanced concepts in the field of intelligent data analysis
2. Understand the main architectures of deep neural networks, namely convolutional networks, recurrent networks and transformers.
3. Develop deep neural networks for application to practical problems
4. Understand the main characteristics of reinforcement learning and understand the situations in which it should be applied
5. Use machine learning tools for the development, training and validation of data analysis models

1. Deep Neural Networks
2. Convolutional networks for computer vision
3. Recurrent networks
4. Natural language processing
5. Transformers
6. Reinforcement learning
7. Generative models
8. Application of TensorFlow and Keras in the development of deep neural networks

NAO