Author(s): Putta Srivani, G Bhoomika, J Srivigna and M Parimala Sai

Abstract: Wireless sensor networks (WSNs) still face the dual problems of energy consumption and data security. Hence, one of the security responsibilities of WSN networks is to prevent them against Distributed Denial of Service (DDoS) and Denial of Service (DoS). Machine learning-based systems are now the only practical option for defending against these kinds of attacks, given traditional packet deep scan methods depend on open field inspection of transport layer security packets and open field encryption is becoming more popular. This research adds to the existing literature by evaluating the impact of machine learning algorithms on the lifetime of WSN networks and the traffic that flows through their nodes. We used a WSN-dataset of varying sizes to assess the performance metrics of various machine learning classification categories, including K-Nearest Neighbour (KNN), Logistic Regression (LR), Support Vector Machine (SVM), Gboost, Decision Tree (DT), Naïve Bayes, LSTM, and Multi-Layer Perceptron (MLP). Results demonstrated that logical and statistical classification categories outperformed others on numerical statistical datasets, and that, across all performance criteria, the Gboost algorithm outperformed the competition. When doing these validations, the following performance measures were considered: accuracy, F1-score, FPR, FNR, and training execution time. In addition, the accuracy, F1-score, FPR, and FNR scores for the Gboost algorithm were 99.6%, 98.8%, 0.4%, and 0.13%, respectively, according to the test findings. In terms of training execution time, it averaged out all datasets at 1.41 seconds. Also, this article proved that numerical statistical data works best with datasets between 3,000 and 6,000 records in size, with at least 50% overlap between each category and all others. In addition, this article examined how Gboost affected the lifespan of WSNs, which was shown to be 32% shorter than other Gboost-free cases.

DOI: 10.33545/26633582.2024.v6.i2b.135

Pages: 125-130 | Views: 668 | Downloads: 290

Download Full Article: Click Here