Log Training

  • Model span (days):
  • Feature vector length:
  • Maximum training data sample value per model:  If the difference of Total Entry and Detection Index exceeds the number, it will trigger the detection (Default: 10000)
  • KeyWord Setting: It allows the project to collect keywords information to do pre-index so that it can have fast query on keywords searching. It can be disabled to improve the detection speed and storage space saving. (Default: Letters and Numbers)
  • Letters and Numbers: it will collect all keywords, except pure numeric ones. (Alert and Deployment project will keep pure numeric keyword)
    • Letters only: it will only collect all alphabetic keywords. If the keywords contain numbers, it will skip.
    • None: disable the keywords collection.
    • Holistic Model:
  • Enable NLP: Extract the feature key phrase for each pattern using the Natural Language Processing. It can be disabled to improve the analysis performance. (Default: OFF)