•  
  •  
 

Article Type

Original Study

Abstract

The spread of wireless networks has led to an increase in serious cyber attacks due to their weak architecture. This article focuses on reevaluating cybersecurity in wireless network technology by integrating statistical information detection methods and artificial intelligence (AI) algorithms. To construct a wireless networking scenario that accurately reflects real-life conditions, we created a data fabrication that included four pre-existing anomalies as well as four newly introduced anomalies. The synthetic dataset created from these generation processes contains 20 thousand distinguishable values, which are later divided into training and validation sets. Using the strategy described before, we began to analyze the data using exponential smoothing. This statistical method is used to detect anomalies in time series data. By using a specific contamination threshold and plotting residuals, the approach adheres to expected trend lines that display variations. In addition, we have integrated an independent Support Vector Machine (SVM), a machine learning technique, to enhance anomaly detection capabilities in different scenarios. The degree of freedom parameter was determined based on the fit to the anomalous data, which subsequently affected the model performance. The research focused on evaluating the effectiveness of the proposed approach. To do this, we used multiple performance metrics, including accuracy, false positive rates, and detection rates. The results showed the effectiveness of the statistical anomaly detection method based on artificial intelligence in accurately identifying cyber threats in wireless networks and mitigating their effects. The one-class SVM classifier achieved a precision of 0 and recall of 1 on the validation set, specifically 0.099. The confusion matrix provides limited insights regardless of whether the model produces 0000 and an F1 score of 0. 1769. It has effectively demonstrated its functionality by prioritizing early detection of defective processes and cells. This measure has improved the reliability of the wireless network. Thus, this study demonstrates the tremendous potential of using statistical anomaly detection techniques, neural networks, and deep learning algorithms to address emerging risks associated with the increasing complexity of wireless networks. The available architecture can serve as a highly effective tool for network administrators and security personnel to protect critical infrastructure and sensitive data from exposure in the digital age.

Keywords

Super vector machine, Machine learning, Artificial intelligence

Download

Share

COinS