Optimal Network Route Estimator Using Prediction Algorithms

Mobile networks have undergone a boost in the recent years, which has posed a big challenge to accurately predict to help in resource usage optimization , enhance the quality of service and save resources. Machine learning (ML) is profoundly used to extract the hidden deep features in the data which is helpful for understanding network traffic instability, but the performance of a only one ML model isn’t quite efficient. The objective of this paper is to design a model using prediction algorithms to decide the best optimal path for network route to aid the efficient transmission of data from source IP address to the destination IP address. This will increase the efficiency of transmission based on quality of service parameters. In this paper, a Machine Learning algorithm called Random Forest is applied to train the prediction model. The algorithm is implemented by using python and its enriched library resources which include Pandas to load the data, NumPy to reshape the data, Matplotlib to plot the graphs, SciPy for advanced computation and SkLearn for data mining and data analysis. Real-life traffic dataset is fed into the model for the evaluation. Through experimental results, it is found that compared to other single ML algorithms the proposed model is much more efficient and improves the prediction performance. As per the information and papers available on the internet, the maximum achieved accuracy is found to be about 92% using XGBoost Model. During the initial model testing, the accuracy obtained was 72% using Random Forest Algorithm. Including other parameters and changing the model accordingly, the accuracy gradually increased to 98%.