Proposta de um método de análise do desempenho de redes móveis com base na representação dimensional de KPIs usando PCA e clusterização
Felipe A M Correa, Alan Pinheiro, Willian Nunes

DOI: 10.14209/sbrt.2019.1570550637
Evento: XXXVII Simpósio Brasileiro de Telecomunicações e Processamento de Sinais (SBrT2019)
The complexity, coverage and quantity of data produced by mobile networks have required new methods for a more automated analysis and able to search for patterns of non-obvious anomalies. In this sense, this research suggests a method of detection of anomalies based on reduction of dimensionality, clustering of performance indicators and segmentation in different periods of the day. The method was evaluated using six indicators of a network in operation and the results found encourage the use of the method as a tool to guide the correction of failures, optimize resources and indication of expansion of network capacity.