- Problem statement
- Describe the theoretical and practical relationships among the following machine learning methods:
- Full Bayesian approach
- Maximum a posteriori estimation approach (MAP)
- Maximum likelihood estimation approach (ML)
- Backpropagation neural network approach (BP)
- Also, discuss the implications of the above methods with respect to the following principles for machine learning:
- Occam's razor principle
- Maximum entropy (ME) principle
- Minimum description length (MDL) principle
- The report does not have to contain any original contributions of the authors, but should be based on a broad survey of relevant references.
- The report must be written in English using LaTex. The length should be
between 5-10 pages using the style file available here.