Machine Learning Based Cooperative Spectrum Sensing Using Regression Methods

New cooperative spectrum sensing (CSS), it encounters better spectrum efficiency, has come into existence as a new strategy. In this study, techniques based on machine learning (ML) work together with CSS to improve user understanding; this is only possible when ML algorithms predict channel states. Additionally, many popular regressions machine learning models such as linear regression, nonlinear regression, generalized linear model, regression tree, and support vector machine regression (SVM) ensembles are discussed along with statistical analysis. This article assumes that 20 main users have different power, controls the maximum power of 20 units, and fixes the threshold of 4 units for all four cases, and analyses the results. In this respect, we came to the conclusion that a best fit line should be used for obtaining characteristics such that linear regression could be used. To further enhance regression and achieve the best results, other changes were also employed.