Anomaly detection in a network based on a key performance indicator prediction model
摘要
說明書
Additionally, or alternatively, the network monitoring platform may train the KPI prediction model using a supervised training procedure that includes receiving input to the KPI prediction model from a subject matter expert, which may reduce an amount of time, an amount of processing resources, and/or the like to train the KPI prediction model relative to an unsupervised training procedure. In some implementations, the network monitoring platform may use one or more other model training techniques, such as a neural network technique, a latent semantic indexing technique, and/or the like. For example, the network monitoring platform may perform an artificial neural network processing technique (e.g., using a two-layer feedforward neural network architecture, a three-layer feedforward neural network architecture, and/or the like) to perform pattern recognition with regard to patterns of whether parameters associated with historical values increases or decreases a value of the prediction of the KPI, or does not increase or decrease the value of the prediction of the KPI (with a threshold confidence of correlation), and/or the like. In this case, using the artificial neural network processing technique may improve an accuracy of the KPI prediction model generated by the network monitoring platform by being more robust to noisy, imprecise, or incomplete data, and by enabling the network monitoring platform to detect patterns and/or trends undetectable to human analysts or systems using less complex techniques.
