Extracting Knowledge from Images of Meanders and Spirals in the Diagnosis of Patients with Parkinson’s Disease

Статья в журнале

In this paper, the problem of diagnosing Parkinson’s disease based on handwritten drawings is investigated. Human-drawn spirals and meanders are used as images. A genetic fuzzy classifier is used as a diagnostic tool. This classifier is built using machine learning methods based on a discrete genetic algorithm. The multiobjective non-dominated sorting genetic algorithm was applied in the work. The diagnostic error, the number of terms, and the number of rules were used as objectives. Higher diagnostic accuracy was achieved compared to methods such as naive Bayes classifier, support vector machine and optimum-path forest. In addition, the fuzzy classifier extracts knowledge for the clinician, which makes it possible to understand causal relationships when making a diagnosis. This is achieved thanks to fuzzy rules of the IF-THEN type. These rules use the fuzzy terms “Small”, “Medium”, “Large” to evaluate the values of the features of the image.

Журнал:

• Pattern Recognition and Image Analysis
• Pleiades Publishing (New York City)
• Индексируется в Scopus

Библиографическая запись: Sarin, K. Extracting Knowledge from Images of Meanders and Spirals in the Diagnosis of Patients with Parkinson’s Disease [Electronic resource] / K. Sarin, I. Hodashinsky, M. Svetlakov // Pattern Recognition and Image Analysis. – 2022. – Vol. 32. – № 3. – P. 658-664. – DOI 10.1134/S1054661822030385

Индексируется в:

• Scopus ( https://www.scopus.com/record/display.uri?eid=2-s2.0-85140243722&origin=resultslist&sort=plf-f )
• РИНЦ ( https://elibrary.ru/item.asp?id=49604139 )
• ВАК ( https://elibrary.ru/item.asp?id=49604139 )