Objectives: With advances in data availability and computing capabilities, artificial intelligence and machine learning technologies have evolved rapidly in recent years. Researchers have taken advantage of these developments in healthcare informatics and created reliable tools to predict or classify diseases using machine learning-based algorithms. To correctly quantify the performance of those algorithms, the standard approach is to use cross-validation, where the algorithm is trained on a training set, and its performance is measured on a validation set. Both datasets should be subject-independent to simulate the expected behavior of a clinical study. This study compares two cross-validation strategies, the subject-wise and the record-wise techniques; the subject-wise strategy correctly mimics the process of a clinical study, while the record-wise strategy does not.Methods: We started by creating a dataset of smartphone audio recordings of subjects diagnosed with and without Parkinson’s disease. This dataset was then divided into training and holdout sets using subject-wise and the record-wise divisions. The training set was used to measure the performance of two classifiers (support vector machine and random forest) to compare six cross-validation techniques that simulated either the subject-wise process or the record-wise process. The holdout set was used to calculate the true error of the classifiers.Results: The record-wise division and the record-wise cross-validation techniques overestimated the performance of the classifiers and underestimated the classification error.Conclusions: In a diagnostic scenario, the subject-wise technique is the proper way of estimating a model’s performance, and record-wise techniques should be avoided.

In this paper, we wanted to discriminate between two groups of patients (patients who suffer from Parkinson's disease and patients who suffer from other neurological disorders). We collected a variety of voice samples from 50 subjects using different recording devices in different conditions. Subsequently, we analyzed and extracted features from these samples using three different Cepstral techniques; Mel frequency cepstral coefficients (MFCC), perceptual linear prediction (PLP), and ReAlitive SpecTrAl PLP (RASTA-PLP). For classification we used leave one subject out validation scheme along with five different supervised learning classifiers. The best obtained result was 90% using the first 11 coefficients of the PLP and linear SVM kernels.

Parkinson's disease (PD) is a neurodegenerative disorder of unknown etiology. PD patients suffer from hypokinetic dysarthria, which manifests on all aspects of voice production, respiration, phonation, articulation, nasality and prosody. To evaluate these disorders, clinicians have adopted perceptual methods, based on acoustic cues, to distinguish the different disease states. To develop the assessment of voice disorders for detecting patients with Parkinson's disease (PD), we have used a PD dataset of 34 sustained vowel / a /, from 34 people including 17 PD patients. We then extracted from 1 to 20 coefficients of the Mel Frequency Cepstral Coefficients from each person. To extract the voiceprint from each voice sample, we compressed the frames by calculating their average value. For classification, we used Leave-One-Subject-Out validation-scheme along with the Support Vector Machines with its different types of kernels. The best classification accuracy achieved was 91.17% using the first 12 coefficients of the MFCC by Linear kernels SVM.