Unsupervised strategies in detecting log anomalies using AIOps monitoring to amplify performance by PCA and ANN systems

A fundamental task that artificial intelligent operations (AIOps) perform is to mitigate the risk of abnormal system behaviours, identify and demystify the alerts when encountering the presence of log anomalies, and assess the reasons for the different system failures and run smoothly. System flaws must be fixed and to empower this functionality, the infusion of related artificial intelligence needs to be integrated. There have been several innovative strategies that have been incorporated with systems utilising AIOps platforms. However, the study has been limited, and some grey areas remain. Suppressing incorrect logs in system performance analysis is unsupervised in this paper. PCA and ANN produce a feed input for detailed analysis. System performance improves. 'Pseudo positives' - false alerts in log anomaly detection theories - are introduced in the study. The proposed strategy reduces aberrant logs by 72%, outperforming most other experiments. It is unique in log analysis since it reduces false positives, making it easier to find true anomalies and improving system efficiency. This approach has promising research possibilities.