Evolution of Machine Learning Techniques for Optimizing Delay Tolerant Routing

This paper discusses about the different Delay Tolerant routing protocols like SAW, Epidemic and PRoPHET, Evaluation metrics for DTN, multilabel class prediction or classification. We have made a survey of Machine Learning involvement in optimizing Delay Tolerant routing. Different classifiers like Naïve Bayes, k-NN, Boosting (XGBoost and AdaBoost), Random forest (Bagging) are discussed which are capable of producing good results in optimizing delay tolerant routing. In the end we have mentioned about the performance metrics that are used to find the accuracy and loss for machine learning techniques. The metrics used are Jaccard similarity score, Zero-one loss, F-1 score and hamming loss.