Learning algorithms and statistical software, with applications to bioinformatics.

Summary Statistical learning is the field of mathematics that deals with the development of data analysis algorithms. This thesis is divided into two parts: the introduction of mathematical models and the implementation of software tools. In the first part, I present new algorithms for segmentation and for data partitioning (clustering). Data partitioning and segmentation are analysis methods that look for structures in data. I present the following contributions, highlighting the applications to bioinformatics. In the second part, I present my contributions to the free software for statistics, which is used for the daily analysis of the statistician.