Granular Network Traffic Classification on Mobile Application Service for Better Parental Control

The widespread use of mobile applications among children had its limitations in excessive screen usage in entertainment. Traditional parental control techniques mostly depend on blocking applications based on their name or categories and are ineffective at identifying potentially new dangerous applications. This study addresses this issue by introducing a paradigm shift for protecting kids' digital experiences through granular network traffic classifications. This system classifies mobile app network traffic into various groups by implementing machine learning or deep learning models. Machine learning models like linear regression, Adaboost Classifier, Support Vector Classification (SVC), Ridge Classifier and deep learning models like Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), and Long Short Term Memory (LSTM) were used to identify and classify mobile applications. Initially, the deep learning RNN model results had 80% accuracy in the classification of mobile apps, and the machine learning logistic SVC and Adaboost Classifier had 72% accuracy. After adding new applications to the existing dataset, the accuracy of the RNN model has remained the same 80% accuracy. Meanwhile, the machine learning model accuracy reduced to 70%, proving the flexibility of the system. This work can assist parents in monitoring their kids to use mobile applications more for education. This work has a huge scope for information technology management of social change using hybrid learning models in software-defined networking in local networks for the administrators to empower their users to wisely use the internet.