USING MACHINE LEARNING METHODS TO CREATE AN ALGORITHM FOR ADAPTIVE ESTIMATION OF PROJECT TASK COMPLETION TIME

This scientific article explores the application of machine learning methods for estimating project task completion time. The authors analyze the limitations of traditional methods, such as the PERT diagram and the CPM method, and identify their fuzziness and inadequate accuracy in complex and dynamically changing projects.The article presents various machine learning approaches to time estimation, including regression analysis, convolutional neural networks (CNN), and decision trees using random forest. Each approach has its advantages and disadvantages, such as the ability to handle large volumes of data, capture complex dependencies, and interpretability of models.The authors also introduce their algorithm for adaptive estimation of task completion time based on machine learning methods. The algorithm undergoes testing on different projects, demonstrating significant improvement in forecast accuracy compared to traditional methods. Training on historical data, task duration forecasting, and retraining based on results are key components of the algorithm.The research findings confirm that the use of machine learning methods significantly enhances the accuracy of project task completion time forecasts. This is of great importance for effective project management, informed decision-making, and achieving more precise timelines