Creating Next Gen Log Analysis with AI - DRAFT

Stage 2 - Use TensorFlow / Keras based AI and ML to detect patterns in data. We used K-Nearest neighbors algorithm to identify and classify patterns. Stage 3 - This is people centric stage where the output from Stage 2 is consulted with BI / Admins and System admins along with Business Stewards to help identify which type of errors effect organizations more (this stage ensures that classification priority of errors are organization specific and not generic). Stage 4 - Learn now again from the tags in Stage 3 and build and distribute models. We used SVM model this time to classify errors as severity 1-5 (target label 0-4 in multi-class classification).