Data analysis and pattern recognition

The purpose of data analysis and pattern recognition is to analyse and make explicit the concepts embedded in large amounts of data that can come from many sources. These methods have ever-increasing application benefits in fields as diverse and varied as computer vision, signal analysis, robotics, medicine, finance, electronic commerce, or military applications, etc. This course therefore aims to introduce the fundamental principles and techniques of data analysis and pattern recognition, and in particular descriptive approaches (automatic description of the concepts contained in the data), as well as predictive approaches.

• Factor Analysis (PCA, AFC, ACM)
• Unsupervised classification (HAC, Kmeans)
• Linear models for regression
• Logistic regression for classification
• Problem of over-fitting and regularization
• Neural networks: representation and learning
• Tips and Practices for Applying Machine Learning
• Reinforcement learning

Final mark = 50% Knowledge + 50% Know-how Knowledge mark = 100% final exam Know-how mark = 100% continuous assessment