Anomaly detection: Difference between revisions

From LemonWiki共筆
Jump to navigation Jump to search
No edit summary
Line 7: Line 7:
* All values is event or odd
* All values is event or odd
* The values are the same even the column is totally different
* The values are the same even the column is totally different
* For consumer data
** Season issue: consumption data of coat (大衣) and cold weather (winter 冬天)
** Holiday issue: consumption data of special holiday e.g. Mid-Autumn Festival / Moon Festival


== Anomaly detection of categorical data (qualitative variable) ==
== Anomaly detection of categorical data (qualitative variable) ==
Line 18: Line 15:
* Dramatically Increase or decrease of rows count for each time period
* Dramatically Increase or decrease of rows count for each time period


== Anomaly detection for stirng data ==
== Anomaly detection for consumer data ==
* For consumer data
** Season issue: consumption data of coat (大衣) and cold weather (winter 冬天)
** Holiday issue: consumption data of special holiday e.g. Mid-Autumn Festival / Moon Festival
 
== Anomaly detection for string data ==


* created time of the text message
* created time of the text message

Revision as of 15:44, 3 October 2022

Outlier / Anomaly detection

Anomaly detection of numeric data

  • Median
  • Range Checks
  • All values is event or odd
  • The values are the same even the column is totally different

Anomaly detection of categorical data (qualitative variable)

  • Normal distribution e.g. The interest of audiences should be very different NOT coherent

Anomaly detection for time series data

  • Trend
  • Dramatically Increase or decrease of rows count for each time period

Anomaly detection for consumer data

  • For consumer data
    • Season issue: consumption data of coat (大衣) and cold weather (winter 冬天)
    • Holiday issue: consumption data of special holiday e.g. Mid-Autumn Festival / Moon Festival

Anomaly detection for string data

  • created time of the text message
  • time frequency of the text message
  • length of the text message

More on: Outlier - Wikipedia