Selección de características de la voz de personas con Parkinson, usando algoritmos genéticos

Abstract

Nowadays, Parkinson's disease (PD) is one of the neurodegenerative diseases that affects most around the world, just behind Alzheimer's disease. There are about 6.2 million people suffering this disease and it is estimated that in 2040 the number of patients with PD will double. Even when we know the importance of this disease, in our country does not exist a conscience and a culture about it and in consequence there is not an exact data about people who suffer from this ailment. Symptoms of PD include slow movement (bradykinesia), stiffness, tremor, postural instability, decreased facial expressions among others, but it is often confused with patient's age as part of physiological change of them. This disease does not have a definitive cure, even after the surgery that some patients could take in order to reduce the symptoms of PD, the rest of patients must continue with the pharmacologic treatment. This technical study focuses on select the features of the voice that get more incident in the diagnostic of the disease. For this, we are going to use the genetic algorithms in the MATLAB environment, and then, we are going to realize a comparison with other results from different methods of feature selection.