DOI: 10.17587/prin.13.187-199

Applying Unsupervised Machine Learning Algorithms to Ensure Requirements Consistency

K. I. Gaydamaka, k.gaydamaka@gmail.com, Department of Systems Engineering, MIREA — Russian Technological University, Moscow, 119454, Russian Federation, А. D. Belonogova, alena.belonogova@yandex.ru, National Research Nuclear University MEPhI, Moscow, 115409, Russian Federation

Corresponding author: Gaydamaka Kirill I., Postgraduate Student, Department of Systems Engineering, "MIREA — Russian Technological University", Moscow, 119454, Russian Federation, E-mail: k.gaydamaka@gmail.com

The article is devoted to the problem of ensuring the quality of requirements for complex technical systems. The purpose of this article is to apply unsupervised machine learning techniques to test a set of requirements for consistency. It is assumed that the clustering of requirements will allow us to determine the requirements that are closest in meaning to the given one, and this may indicate a possible contradiction and require additional check of potentially conflicting requirements. The article discusses a comparison of clustering methods such as k-means, agglomerative hierarchical clustering, DBSCAN, EM-algorithm, as well as methods for converting sentences into numeric vectors TF-IDF, doc2vec, BERT. BERT and k-means showed the best result — a cluster that included only conflicting requirements.

Keywords: natural language processing, requirements clustering, requirements management, requirements engineering, requirements consistency

pp. 187—199

For citation:
Gaydamaka K. I., Belonogova A. D. Applying Unsupervised Machine Learning Algorithms to Ensure Requirements Consistency, Programmnaya Ingeneria, 2022, vol. 13, no. 4, pp. 187—199.