Search

The classification of wireless communication channel scenarios is vital for modern wireless technologies. Efficient data preprocessing for identification, especially starting from 5G and beyond, where multiple scenario transitions occur, is crucial. Machine Learning (ML) is employed for scenario identification. Moreover, accurate ML classification is required to enhance the decision-making process in each communication layer. The proposed model in this study utilizes an enhanced preprocessing phase. The proposed model proves that adding the variance inflation factor (VIF) elimination layer has a significant effect in eliminating the residual noise after regularization. By evaluating the VIF, the high multi-collinear features are removed after adding a regularization penalty. Consequently, the total explained variance (TEV) was enhanced by 5% and reached 76%. Thus, the classification accuracy of the identification processes of different rural and urban scenarios was increased by 3%, on average, compared with previous work for each algorithm: Random Forest (RF), K-Nearest Neighbor (KNN), Support Vector Machine (SVM), and Gaussian Mixture model (GMM). [ABSTRACT FROM AUTHOR]

Wireless communication channel scenario classification is crucial for new modern wireless technologies. Reducing the time consumed by the data preprocessing phase for such identification is also essential, especially for multiple-scenario transitions in 6G. Machine learning (ML) has been used for scenario identification tasks. In this paper, the least absolute shrinkage and selection operator (LASSO) is used instead of ElasticNet in order to reduce the computational time of data preprocessing for ML. Moreover, the computational time and performance of different ML models are evaluated based on a regularization technique. The obtained results reveal that the LASSO operator achieves the same feature selection performance as ElasticNet; however, the LASSO operator consumes less computational time. The achieved run time of LASSO is 0.33 s, while the ElasticNet corresponding value is 0.67 s. The identification for each specific class for K-Nearest Neighbor (KNN), Support Vector Machine (SVM), and k-Means and Gaussian Mixture Model (GMM) is evaluated using Receiver Operating Characteristics (ROC) curves and Area Under the Curve (AUC) scores. The KNN algorithm has the highest class-average AUC score at 0.998, compared to SVM, k-Means, and GMM with values of 0.994, 0.983, and 0.989, respectively. The GMM is the fastest algorithm among others, having the lowest classification time at 0.087 s, compared to SVM, k-Means, and GMM with values of 0.155, 0.26, and 0.087, respectively. [ABSTRACT FROM AUTHOR]