Performance Enhancement to Improve Accuracy for the Novel CNN GoogleNet Compared AAgainst ResNet in Predicting Cyber Hacking Breaches

The aim of this research is to predict cyber hacking breaches by using the two machine learning algorithms Novel CNN Google Net and ResNet algorithms. Materials and Methods: Novel CNN Google Net and ResNet are used in this work to predict the cyber hacking breaches with varying training and testing splits. There are 2030 cyber hacking breach images collected for this research study from an open source Kaggle web resource. The datasets were split into two train dataset and test dataset. The train data set consists of 1624 images and test data consists of 406 data set images. The information has been gathered from various web resources. The G-Power test utilized approximately 80% of the G-Power setting parameters with
$\boldsymbol{\alpha}=\mathbf{0.05}$
and power=0.80. The experiment study was iterated twenty times using the above said models. Result and discussion: This research paper found the Novel CNN Google Net 93.40% has the increased accuracy over ResNet 92.05%. The Independent Sample T-test showed statistical significance at the 95% confidence interval with a p-value of 0.001
$(\mathbf{p} < \mathbf{0.05})$
. This indicates a significant difference between the two groups. Conclusion: Machine learning algorithms are used in the study paper. both. Novel CNN GoogleNet and the ResNet algorithm. Accuracy comparison of GoogleNet compared over ResNet in predicting cyber hacking breaches in the early stage.