ALL DOCUMENTS ARE IN FRENCH
The most important part of this project is the report found in report.pdf that answers every questions asked in the project description project.pdf.
This academic classification project was included in a course at the UTC in collaboration with CapucinePr.
It deals with data mining and machine learning.
The purposes of this project were to :
- Test in R different classification methods such as discriminant analysis, decision tree, logistic regressoin on a set of heteregeneous datasets.
- Theoretically justify a new discrete discriminant analysis method.
- Implement it in R.
- Compare for each dataset the results of classification methods
- Interpret every result with regards to machine learning knowledge taught during the course : shape of the data, number of variables, ...
The R scripts included in this repository were changed constantly and are not meant to be unique readable clearly structured file but may give an idea of the iterative process we went through to test and implement various classification methods.