A notable problem in the contemporary expanding networking paradigm is the congestion resulting from the substantial volume of data transmitted between nodes in any network. During congestion, packets may become damaged or lost. By evaluating the relationships among various network QoS measures, we may identify and address congestion issues that occur in such scenarios. This study use singular value decomposition (SVD) to mitigate congestion issues. Singular value decomposition (SVD) is a matrix factorization method utilized in diverse applications, such as principal component analysis (PCA) and linear regression. We employ an innovative method known as SVDULR, which utilizes singular value decomposition and linear regression to efficiently identify congestion and improve service quality.