<![CDATA[General Background Network data traffic rating represents a core task in modern network engineering for service quality control and bandwidth management. Specific Background Traditional approaches relying on deep packet inspection and port-based categorization face severe limitations due to the proliferation of dynamic ports and widespread encryption protocols such as TLS and HTTPS. Knowledge Gap Although automated algorithms offer alternative solutions, comprehensive applications detailing complete processing pipelines that systematically compare multiple predictive models without inspecting actual packet contents remain scarce. Aims This study designs an integrated flow-based data processing pipeline to extract nine statistical properties for categorizing six data movement classes using four distinct automated algorithms. Results Utilizing a simulated dataset of 3000 flows, experimental outcomes revealed that the aggregate tree model attained the highest precision of 97.87 percent and a macro F1-score of 97.87 percent, significantly outperforming Decision Tree, Support Vector Machine, and KNN models. Novelty This research demonstrates the explicit superiority of ensemble techniques in data categorization by identifying byte rate, mean packet length, and total bytes as the most decisive flow-level metrics, entirely independent of payload decryption. Implications Relying on extracted statistical properties presents a highly reliable methodology for managing modern encrypted infrastructures, suggesting that administrators should prioritize aggregate machine learning systems to optimize resource allocation and detect unwanted activities. Highlights: Aggregate tree models attained a 97.87 percent precision rate across six data movement categories. Payload decryption is unnecessary when utilizing statistical properties such as byte rates and packet lengths. Machine learning pipelines successfully categorized video streaming, web browsing, and voice communications without port dependency. Keywords: Machine Learning, Data Movement Categorization, Statistical Properties, Deep Packet Inspection, Decision Tree Models]]>
