Titova, Vera YuriivnaТітова, В.Ю.2019-01-212019-01-212018Titova, V.Y. Comparative analysis of neural networks for the evaluation of software quality [Текст] / V. Y. Titova // Вісник Хмельницького національного університету. Технічні науки. – 2018. – № 5. – C. 7-12.https://elar.khmnu.edu.ua/handle/123456789/7278The problem of software quality evaluation is considered in this paper. Software quality is evaluated using a quality model. The quality model consists of software quality metrics classified into a hierarchical tree structure. The upper level of this structure consists of quality characteristics, and the lower level consists of software quality attributes. Based on the analysis of these characteristics and attributes, the authors determine that current quality model is not formalized. So, they propose the formalized model of software quality. This model is the basis for an unsupervised neural network for software quality evaluation. Based on the comparative analysis of clustering validation indexes a Kohonen SOM is chosen. The model and the neural network developed in this paper become the basis for developing a software quality evaluation system.enKohonen SOMART-networkClustering Validation IndexesSoftware Quality ModelNeural NetworksUnsupervised LearningComparative AnalysisАРТ-мережаіндекси перевірки кластеризаціїмодель якості програмного забезпеченнянейронні мережінеконтрольоване навчанняпорівняльний аналізСтаття681.3