Medical equipment predictive maintenance is essential to maintaining consistent and dependable healthcare services. Wireless Sensor Networks are essential for keeping an eye on medical devices and anticipating malfunctions before they happen. In this work, we provide a predictive maintenance strategy for medical WSNs based on LSTMs and use Local Interpretable Model-Agnostic Explanations to improve its interpretability. Our method increases the accuracy of fault predictions while providing decision-making transparency. Results from experiments show how well our model works in real-time to explain contributing elements and spot possible failures..

Wireless Sensor Networks are becoming more and more crucial to the advancement of numerous technologies, particularly when combined with Big Data platforms. Although this connection has a lot of potential, there are challenging security challenges as well. Even though WSNs have been the subject of a lot of research, the security requirements for WSNs functioning in Big Data environments have not yet been thoroughly examined. It is also a crucial use of IoT, allowing sensors to exchange a variety of data. However, because of its inherent unreliability and natural surroundings, such a network is susceptible to numerous types of attacks, including insider attacks. Intrusion detection systems (IDSs) are commonly used in WSNs to protect against insider assaults by putting in place the right procedures and techniques. However, sensors may produce too much data in the big data era, which could reduce the efficiency of WSN computing. An overview of the security concerns and difficulties facing WSNs in the big data era is provided in this study. In order to improve the detection of insider threats and the overall security posture of WSNs, a literature review on cybersecurity IDS on WSN in the context of big data is finally suggested. It highlights advancements in IDS methodologies, including federated learning, machine learning, deep learning, and big data techniques.

In the era of Big Data, where data are continuously and exponentially generated, secure collection of massive volumes of data is essential for enabling effective analysis using advanced techniques such as machine learning. These methods facilitate extracting meaningful insights from structured, semi-structured, or unstructured data generated by websites, devices, and sensors at high velocities. Securing this data, analyzed by modern machine learning techniques (e.g., linear regression) and subjected to continuous streams, presents a significant challenge due to its volume and diverse formats. Traditional security and privacy methods prove inadequate in addressing these complex challenges. As a result, the security and efficiency issue for big data collection still deserves research. This paper proposes a secure mechanism for big data collection offering an unprecedented opportunity for predictive analysis and decision-making for improved security performance and efficiency. Differential confidentiality emerges as a promising solution for business data and confidential data collection. The collected big data is stored securely using DiffPrivLib. The discussion and performance evaluation results show the security and efficiency of the proposed secure mechanism. With the implementation of differential confidentiality on the collected data, the results of the data analysis presented an accuracy of 0.91 which shows a very important efficiency of the system to classify correctly the instances.

Strong security measures are now more important than ever due to the quick growth of wireless sensor networks in sectors like smart cities, healthcare, and agriculture. In order to protect these networks from malevolent assaults, intrusion detection systems (IDS) are essential. However, network operators find it challenging to comprehend the system’s decision-making process due to the lack of transparency in traditional IDS. To address this issue, we propose an explainable artificial intelligence based intrusion detection method for large-scale WSNs. The goal of this research is to combine machine learning models with explain-ability frameworks like SHAP and LIME in order to efficiently identify and clarify anomalies in network data. In order to detect different kinds of attacks, we train and assess a variety of machine learning classifiers, such as decision trees, support vector machines, and ensemble approaches, using the popular KDD Cup 99 dataset. The XIA frameworkS assist discover important aspects that aid in attack detection and offer insights into the model’s decision-making process. Additionally, the model’s durability and scalability make it a feasible option for implementation in extensive WSNs, particularly in agriculture, where network security is essential to guaranteeing the dependability of IoT-based monitoring systems. The results have demonstrate the efficacy of the XAI-based intrusion detection system in attaining high accuracy, precision, recall, and F1-score while maintaining decision-making transparency. This method not only enhances intrusion detection but also provides network administrators with the means to decipher and verify the system’s predictions, increasing confidence in automated security systems.

Managing and processing BigData in geo-distributed datacenters gain much attention in recent years. Despite the increasing attention on this topic, most efforts have been focused on user-centric solutions, and unfortunately much less on the difficulties encountered by Cloud providers to improve their profits. Highly efficient framework for geo-distributed BigData processing in cloud federation environment is a crucial solution to maximize profit of the cloud providers. The objective of this paper is to maximize the profit for cloud providers by minimizing costs and penalty. This work proposes to transfer compute (computations) to geo-distributed data and outsourcing only the desired data to idles resources of federated clouds in order to minimize job costs; and proposes a jobs reordering dynamic approach to minimize the penalties costs. The performance evaluation proves that our proposed algorithm can maximize profit, reduce the MapReduce jobs costs and improve utilization of clusters resources.