APPLICATION OF MACHINE LEARNING METHODS FOR CREATING WINDOWS OPERATING SYSTEM UTILITIES

Abstract

In the modern world, Windows continues to remain one of the most popular operating systems for users. However, with prolonged use of this system, people often encounter problems regarding performance and accumulation of unnecessary files, which can affect computer productivity. Traditional system cleaning tools do not always consider individual user needs, which can lead to loss of important data or insufficiently effective cleaning. The growing requirements for accuracy and safety of the cleaning process emphasize the importance of using machine learning methods to automate this process.This work investigates the development and implementation of an intelligent utility for cleaning and optimization of the Windows operating system using machine learning methods based on the ML.NET library. The program is created in C# language with the application of WinUI 3 for building a modern user interface and PowerShell for performing system optimization. Machine learning technology allows effective classification of files by features (type, size, creation date, modification date, etc.) and determines whether they are potentially unnecessary. The developed program offers both standard cleaning functionality by categories and an intelligent cleaning function that minimizes the risk of deleting important data.To evaluate the effectiveness of the proposed solution, a series of experiments on file classification of various categories and comparative analysis with built-in Windows utilities were conducted. The experiments included testing the accuracy of the machine learning model on a dataset containing files of different types and characteristics. The research demonstrates significant advantages of the intelligent approach, which allows automating the system cleaning process and reducing the number of errors.This research has significant practical value for software developers working on tools for cleaning and optimization of operating systems. It demonstrates how modern machine learning methods can improve the accuracy and safety of cleaning, opening the path to more intelligent and convenient utilities for maintaining computer system productivity.