REVIEW OF MODERN TRENDS AND DEVELOPMENT PROSPECTS OF DEVICES FOR INFOCOMMUNICATION SYSTEMS

DOI: 10.31673/2412-4338.2025.038711

Abstract

The article presents a comprehensive approach to applying artificial intelligence tools for predicting and minimizing risks in Scrum IT outsourcing projects. The authors systematize the main types of risks inherent in iterative software development and analyze the effectiveness of using statistical methods, machine learning, and sentiment analysis to improve the accuracy of team performance forecasts. Key Scrum metrics—velocity, commitment accuracy, cycle time, and defect density—are considered as the basis for building analytical risk models. Particular attention is paid to the integration of multi-factor analysis and algorithms such as Random Forest, Gradient Boosting, and LSTM networks for identifying patterns in team dynamics.

A conceptual architecture for a Decision Support System (DSS) is proposed. This system integrates data from Jira, Slack, and GitHub, provides automated collection and processing of metrics, and generates forecasts and recommendations for the Scrum Master in real-time. The DSS encompasses five interconnected components—a data collection layer, a storage repository, an analytical engine, a visualization interface, and a recommendation module— enabling proactive risk management at all stages of the sprint.

The research results prove that the application of AI methods increases the accuracy of delay prediction from 70– 75% (when using classical statistical approaches) to 85–94%, allowing for the timely identification of technical, organizational, and communication risk factors. The practical value of the work lies in the possibility of implementing the DSS approach without developing separate software, by utilizing existing corporate analytics tools. The article forms a theoretical basis for creating intelligent monitoring systems in Agile environments and outlines directions for further research—expanding sentiment analysis for the Ukrainian language, implementing deep learning approaches, and automatically optimizing task allocation.

 Keywords: Scrum, risk management, artificial intelligence, machine learning, decision support system, sentiment analysis, Agile metrics.