Detection of Speech Impairments in Parkinson Disease Using Handcrafted Feature-Based Model on Spanish Speech Corpus

Abstract

Parkinson disease is a neurodestructive disorder. It gradually dismantles the dopamine chemical producing cells in the brain. The symptoms of the disease arise gradually resulting in infects in movements, olfactory, or speech impairments. The cure for disease is important to prevent patients from major motor and non-motor defects. Computer-aided techniques have been introduced for disease detection and is an open research area. In this paper, we present an approach for early diagnosis of Parkinson disease using the Spanish speech dataset. We modeled handcrafted features from different sets of Spanish recordings and classify them using a machine learning model. Several Machine learning algorithms are implemented for significant classification and compared each other’s results graphically. The results depict that handcrafted features of vowels are most efficient in Parkinson disease detection.