Application to regression models.

This example is dedicated to explain regression models and some duality concepts. First the three most common methods, least squares, minimax and least absolute values are described and the last two stated as linear programming problems. The dual of the minimax problem is obtained and its mathematical interpretation is given. Finally, an example of simulated data with three outliers is given and the three methods used to fit the data. The lesson ends with the interpretation of the results of the example.