It is essential and unavoidable to detect Malware on the Internet, as a wide range of online IT services are available. Portable Executable files are the most frequently targeted platform by Malware. Malware must be promptly identified and alerted in a real-world environment by establishing a deployable learning system. The researchers applied machine learning to a Malware dataset, observing the model's performance metrics at a high computational cost, but were unable to deploy the model in a real-world environment. A deployable machine learning model using RF, attaining an accuracy of 97.16%, precision of 95.21%, and F1 score of 95.24% is achieved in the proposed research work, which is particularly adept at accurately identifying Malware. We have developed a novel classification model that employs the Support Vector Machine (SVM) to classify preprocessed data, detecting malware and normal instances. Furthermore, the SHAP technique identifies significant features, including SizeOfStackReserve, DllCharacteristics, and MajorImageVersion. The use of SHAP values facilitates an understanding of the characteristics of each feature in the model's prediction. Employing the SHAP algorithm using the trained SVM model to reduce the features, attained an accuracy of 97.16%.

The increasing ubiquity of Social Internet of Things (SIoT) devices necessitates innovative data aggregation techniques to distill meaningful insights from diverse sources. This study introduces a Dynamic Data Aggregation Model for SIoT devices. The model aims to amalgamate static and mobile device data, employing Density-Based Spatial Clustering of Applications with Noise (DBSCAN) for spatial clustering and Recurrent Neural Networks (RNN) for predicting mobile device movement patterns. The purpose is to offer a holistic approach to predictive analytics in the SIoT domain by seamlessly integrating these methodologies. The model begins with data preprocessing, ensuring data quality. It then applies DBSCAN for spatial clustering, enabling a comprehensive understanding of spatial relationships between devices. Simultaneously, RNNs are used for predictive modeling, specifically in forecasting mobile device movement patterns. The integration of DBSCAN clustering and RNNs forms the model’s innovative core, providing a unified solution for dynamic data aggregation. Comprehensive testing demonstrates the model’s notable accuracy in predicting mobile device movement patterns. By combining the spatial clustering capabilities of DBSCAN with the predictive power of RNNs, the model effectively unifies static and mobile data, advancing predictive analytics in the SIoT context. The proposed model yielded average values of 0.14604 (Mean Squared Error), 2.678385 (Mean Absolute Error), 0.307154 (Root Mean Squared Error), and 1.342317 (Mean Absolute Percentage Error), respectively. The Dynamic Data Aggregation Model proves its efficacy in addressing SIoT challenges. The integration of DBSCAN clustering and RNNs offers a versatile framework for dynamic data analysis, contributing significantly to predictive analytics in SIoT contexts.

The evolution of the Internet of Things (IoT) into the Social Internet of Things (SIoT) involves the integration of social networking features into smart devices. In this paradigm, smart devices emulate human social behavior by forming social relationships with other devices within the network. These relationships are leveraged for service discovery, emphasizing the need for robust security to foster collaboration and cooperation among devices. Security is paramount in the SIoT landscape, as malicious messages from devices can disrupt service functionality, impacting service quality and reliability. These challenges are particularly pronounced in social networks, introducing unique considerations such as heterogeneity and navigability. This study introduces a Security Framework for the Social Internet of Things, adopting a Relativity Strength Approach to enhance the security and reliability of IoT devices within social network contexts. The framework incorporates a relativity-based security model, utilizes Q-learning for efficient device navigation, and employs decision tree classification for assessing service availability. By optimizing hop counts and considering the strength of relationships between devices, the framework enhances security, resource utilization, and service reliability. The proposed security framework introduces a” Relationship key” derived from device-to-device relationships as a central element. This key, coupled with a standard 256-bit Advanced Encryption Standard (AES) algorithm, is employed for encryption and decryption processes. The relationship key technique ensures data protection during transmission, guaranteeing confidentiality and service integrity during network navigation. The system demonstrates an overall security effectiveness of 88.75%, showcasing its robustness in thwarting attacks and preventing unauthorized access. With an impressive overall communication efficiency of 91.75%, the framework minimizes errors and delays, facilitating optimal information trans- mission in smart environments. Furthermore, its 97.5% overall service availability assures a continuous and reliable user experience, establishing the framework’s capability to deliver secure, efficient, and highly accessible smart services.