Contents lists available at ScienceDirect Medical Hypotheses journal homepage: www.elsevier.com/locate/mehy Association analysis of Parkinson disease with vocal change characteristics using multi-objective metaheuristic optimization Elif Varol Altay ⁎ , Bilal Alatas Department of Software Engineering, Firat University, Elazig, Turkey ARTICLE INFO Keywords: Parkinson disease Numerical association rules mining Multi-objective optimization ABSTRACT Parkinson's disease (PD) is a neurodegenerative disorder that has important economic and social effects influ- encing the quality of patient life. Diagnosis of PD is performed in terms of certain criteria depending on the clinical symptom evaluation. However, this method may be inadequate, especially during the onset of the dis- ease. Acoustic analysis of PD is a cost-effective, easy, and non-invasive method for early diagnosis. The mining of association rules is one of the problems in data mining that aims to find valuable and interesting associations in huge data sets. Although association analysis is very popular and useful, to the best of our knowledge, there is not any study on association analysis of PD using vocal change characteristics. Automatic mining of compre- hensible, interesting, and accurate association rules in PD data sets containing huge numerical processed voice data is aimed in this study. Due to the numerical characteristics of the vocal attributes in pre-processed PD data, classical association rules mining methods cannot be efficiently applied to this problem. For this reason; MOPNAR, NICGAR, and QAR_CIP_NSGAII that are artificial intelligence-based algorithms were modeled for mining of numerical association rules in order to obtain better performances without using any pre-process for numerical data for the first time. Furthermore, the problem of association analysis of PD with vocal change characteristics was modeled as a multi-objective optimization problem considering many different com- plementary/contradictory metrics such as support, confidence, comprehensibility, interestingness, etc. in this study. According to the obtained multi-objective rule sets, the NICGAR outperformed in terms of average con- fidence, average CF, average netconf, average yulesQ, and average number of attributes. Introduction Parkinson's disease (PD) is a neurodegenerative disorder caused by damage to neurons responsible for secreting dopamine in the brain, showing major symptoms such as speech disorders (dysphonia), gait disorders, dementia, and tremors [1]. Although some of the symptoms of PD can be decreased by drug therapy or surgical intervention today, there is no biomarker that can be designated by laboratory tests and there is no method for a definitive diagnosis of PD yet. Neurologists diagnose the disease with other diagnostic procedures such as blood tests, neuroimaging techniques, in addition to the physical examination and medical history of patients. Unfortunately, the findings are mis- leading and misdiagnosed as they may show symptoms similar to other neurological disorders such as MSA and PSP in the early stages of the disease. Furthermore, although various drug treatments are used to minimize the difficulties caused by the disease, PD is often diagnosed by invasive methods and this makes the diagnosis and treatment processes of patients suffering from the disease inextricable [2]. The control mechanism of the muscles in the face, mouth, and throat used in the speech formation and speech system in Parkinson's patients is affected by dopamine deficiency. Dysphonia, which char- acterizes the disorders in the formation of sound, is one of the sec- ondary symptoms of the disease, but it is an important parameter that allows distinguishing the sick individuals from healthy individuals. Decreased loudness, roughness, monotonous sound, difficulty in breathing, difficulty in speaking, stress or rhythm disturbance, stut- tering, and slow or rapid speech are common problems affecting ap- proximately 90% of Parkinson's patients [3]. Diagnosis of PD is performed concerning certain criteria depending on the clinical symptom evaluation. However, these clinical symptoms do not occur until dopaminergic neurons loss reaches a level of 60–80% [4]. Recent works claim that when clinical motor symptoms are very mild, speech disturbances occur and even speech analyzes can detect these distortions before a perceptible, pronounced dysarthria occurs [5]. Acoustic analyzes assure early substantial findings without harming the patients. https://doi.org/10.1016/j.mehy.2020.109722 Received 23 March 2020; Received in revised form 8 April 2020; Accepted 8 April 2020 ⁎ Corresponding author. E-mail addresses: evarol@firat.edu.tr (E.V. Altay), balatas@firat.edu.tr (B. Alatas). Medical Hypotheses 141 (2020) 109722 0306-9877/ © 2020 Elsevier Ltd. All rights reserved. T