The rapid advancement of smart devices has resulted in a significant increase in data generation and diversity. Consequently, there is a need for novel network solutions to effectively analyze and comprehend the traffic. In order to effectively manage the substantial volume of data automatically, it is imperative that the proposed solutions exhibit both intelligence and scalability. In order to effectively manage the substantial volume of data automatically, it is imperative that the proposed solutions possess both intelligent and scalable attributes. The advancement of high-performance computing (HPC) has made it increasingly viable to implement machine learning (ML) techniques for addressing intricate problems and its efficacy has been substantiated across several sectors such as healthcare and computer vision. This study has been centered on the analysis of network data with the aim of delineating network slices based on traffic flow patterns. In the context of dimensionality reduction, the process of feature selection has been employed to identify and choose the most pertinent features from a substantial real-world dataset including 80 applications, encompassing over 3 million occurrences. Next, an unsupervised K-Means clustering algorithm is utilized in order to gain a deeper comprehension and differentiation of traffic behaviors. The findings exhibited a strong association among the cases inside the same cluster produced by the unsupervised learning algorithm. A regression technique within the field of supervised machine learning is employed to assess the traffic flow of a network, specifically in terms of byte data. The integration of this technology can be extended inside a practical setting through the use of network function virtualization.