ml and mcdm for abnormal cell detection in 5g & b5g networks

Self-organizing communication networks are a vital pillar in 5g and b5g technology, which operate automatically without human intervention in self-healing, self-configuring, and self-optimizing. self-healing in these networks predicts and resolves network problems and improves performance with the following three methods in the research conducted: rule-based, algorithmic, and machine-learning approaches. this research used the topsis technique as a multi-criteria decision-making method to rank and score cells after data preprocessing. then, based on the rank of each cell, it is divided into two classes: normal and abnormal. then, with three algorithms, decision tree, new bayes and random fars, normal and abnormal cell prediction was performed independently. in the last step, using the combined method of maximum voting, the algorithm was completed and the results showed an improvement in the parameters precisio=0.939, recall=0.962, f-measure=0.968, accuracy=94.0717.