IMP (Instituto Mexicano del Petróleo, Mexican Oil Institute) uses buoys to sensor ocean in order to obtain any kind of data: climatic, waves height and so on. But, since ocean is a very chaotic and changing environment very often sensor cannot take data or they fail to take data out of order.
IMP use data obtained to design contingence plans, construction materials for oceanic stations and in general to monitor challenges against population. Then it is important to deal with missing and "contaminated" data.
In this research I used statistics and artificial intelligence paradigms to infer data tendencies and to make predictions over them.