Machine Learning :classification and evaluation [ LINGI2262 ]

> https://www.icampus.ucl.ac.be/claroline/course/index.php?cid=INGI2262

• Learning as search, inductive bias
• Combinations of decisions
• Loss function minimization, gradient descent
• Performance assessment
• Instance-based learning
• Probabilistic learning
• Unsupervised classification

• understand and apply standard techniques to build computer programs that automatically improve with experience, especially for classification problems
• assess the quality of a learned model for a given task
• assess the relative performance of several learning algorithms
• justify the use of a particular learning algorithm given the nature of the data, the learning problem and a relevant performance measure
• use, adapt and extend learning software

• use the technical documentation to make efficient use of existing packages,
• communicate test results in a short report using graphics.

The 4 mini-projects worth 30 % of the final grade, 70 % for the exam (closed-book).

The mini-projects can NOT be remade in second session
30 % are already set at the end of Q2 and included as such in the final score in the second session.

• Lectures
• Written assignment and/or Miniproject (2 students/group, from 1 to 3 weeks)
• Assignment feedback

• Decision Tree Learning: ID3, C4.5, CART, Random Forests
• Linear Discriminants: Perceptrons, Gradient-Descent and Least-Square Procedures
• Maximal Margin Hyperplanes and Support Vector Machines
• Probability and Statistics in Machine Learning
• Performance Assessment: Hypothesis testing, Comparing Learning Algorithms, ROC analysis
• Gaussian Classifiers, Fisher Linear Discriminants
• Bayesian Learning: ML, MAP, Optimal Classifier, Naive Bayes
• Instance-based learning: k-NN, LVQ
• Clustering Techniques

Required Slides available on:

http://www.icampus.ucl.ac.be/claroline/course/index.php?cid=INGI2262

and more generally all documents (set of mini-projects) available on the same site.

Background:

• LSINF1121 : algorithmics
• LBIR1304 or LFSAB1105 : probability et statistics