<![CDATA[Increasingly evolving digital era, optimizing computer network performance has become a critical priority to support the demands of high-speed, reliable, and adaptive connectivity. This study explores the application of Machine Learning (ML) techniques as a strategic approach to enhancing network performance, with a focus on three key algorithms: Random Forest for detecting anomalies, Long Short-Term Memory (LSTM) for predicting network traffic patterns, and Deep Reinforcement Learning (DRL) for optimizing routing decisions. These algorithms were selected based on their respective strengths in handling complex, high-dimensional network data and their ability to adapt to dynamic conditions. The experimental results demonstrate that the implementation of ML can significantly improve key network performance metrics. Specifically, network throughput increased by up to 21.7%, latency was reduced by 27.1%, and the packet loss rate decreased by 62.2%. These improvements suggest that ML-based approaches offer a promising direction for achieving more intelligent and efficient network management. However, the study also highlights several challenges that must be addressed for broader adoption. These include the dependency on high-quality, real-time data; the computational overhead required for model training and deployment; and potential vulnerabilities related to cybersecurity threats. In addition, the interpretability and transparency of complex ML models remain areas requiring further exploration. Overall, this research contributes to the growing field of ML-driven network optimization and provides a foundation for future advancements in autonomous, adaptive, and secure networking systems. It opens up new possibilities for the development of next-generation communication infrastructures that are smarter, more resilient, and performance-oriented.]]>
