Non-linear regression and input variable importance in Machine Learning with CART and Random Forests

PhD training activity specially aimed at students from the Manufacturing and Industry Doctoral Area

DESCRIPCIÓN

The machine learning tool of this seminar will be the non-parametric techniques of CART (Classification and Regression Trees) and two 'sophistications (of the original CART): Bagging and Random Forests, capable of capturing highly non-linear behavior. The latter provide the interesting added-value of quantifying input variable importance. The methodology would be illustrated by means of examples with real or simulated data sets, using the R software.  The real data examples include bank counterfeit bank note detection, wine quality prediction, flower species classification, housing price prediction, thyroid disorder diagnosis, breast cancer diagnosis, and digit image classification from images.

OBJETIVOS

To provide the students with a Machine Learning tool for the estimation of non-linear input-output relationships in classification (categorical output) or regression (quantitative output) problems, as well as to quantify the importance of the inputs on the output

Fecha y horario: 26 de mayo . De 16:00 a 19:00 o 20:00 h

Presentación: On-line

Importante: Para la inscripción, es necesario utilizar el correo institucional UPM

Email de contacto: doctorado@upm.es