A Multiple Linear Regression Model for Predicting Congestion in Heterogeneous Vehicular Networks

Finite capacity of network resources and enormous data generated by vehicles using safety and comfort applications, have made network congestion a challenge to manage in Heterogeneous Vehicular Network (HetVNET). In this paper, we propose a reliable network congestion model based on a Multiple Linear Regression (MLR), which is a supervised machine learning algorithm to predict network congestion in HetVNET. We have evaluated the performance of our proposed network congestion prediction model using a Cross-Validation test approach. Numerical results show that the proposed linear congestion prediction model is reliable, which can explain and support variability of the response as well. Moreover, we have weighted effectiveness of each considered HetVNET parameters, in association with congestion situation in HetVNET.