Enhancing IoT Software Piracy Detection through Feature Extraction and Deep Learning Optimization

Abstract

The Internet of Things (IoT) connects devices, applications, storage, systems and services. This creates chances for cyberattacks as these systems run continuously in organizations. Currently, software piracy and malware infections are considered a serious risk to IoT security. These threats can harm sensitive data and affect both financial and reputational aspects. This research study introduces an integrated strategy to identify pirated software and malware within IoT networks. In this study deep neural network is integrated with TensorFlow to detect software piracy by examining plagiarism in source code. Techniques such as tokenization and weighting are used to remove unnecessary data and also highlight the role of each token used in assessing source code similarity. The Google Code Jam (GCJ) dataset was used as the basis for testing software piracy detection. At the same time deep convolutional neural network (CNN) was used to detect malware within IoT networks by converting malicious code into color images for visualization. The experiments use malware samples from the Maling dataset. This proposed model shows high classification accuracy as compared to existing techniques.