Eureka! Institutional Repository

This thesis is studying machine learning methods in order to identify human activities performed in daily life. We studied the behavior Machines SVM Support vector with three different cores, a linear, a radial (RBF) and polynomial. For the training and validation of learning engineering algorithms used method cascading rolling fixed-size window. Although human performance in different ways from person to person, we were able to predict satisfactorily perform the same activities from different people