A Novel Framework of Network Packet Loss Detection Using Random Forest Algorithm over Support Vector Machine Learning Algorithms to Improve Accuracy

Abstract

The objective of this paper is to employ a novel framework of network packet loss detection with increased accuracy rate using Novel Random Forest classifier compared to a Support Vector Machine (SVM) Classifier. Materials & Procedures: The data set used in this study utilizes the publicly available Kaggle network traffic data set and UCI machine learning repositories. The sample size of network packet loss detection with improved accuracy rate sample size was 50 (Groups 1 and 2 each had 25 participants), and the computation was performed using a G-power of 0.8, with alpha and beta values of 0.05 and 0.2 and a 95% confidence interval. A network packet loss detection with improved accuracy rate is performed by Random Forest (RF) whereas multiple samples <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(\mathrm{N}=10\mathrm{A})$</tex> SVM were a number of samples <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(\mathrm{N}=10)$</tex> Results: The Novel Random Forest classifier has 93.56% higher accuracy rates in comparison to the accuracy rate of SVM is 89.23%. There exists a statistical significance difference between two groups <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(\mathrm{p}=0.0231;\mathrm{p} &lt; 0.05)$</tex> with confidence interval 95%. Conclusion: Novel Random Forest when compared, a classifier offers more accurate results. to SVM for network packet loss detection with improved accuracy rate.