Anomalous Logline Detection Tool

• Machine problems can be detected in an early phase and solved, preventing failure or misfunction and thus standstill times
• Very short training phase, also after software version change

Input(s):

• log text files
• mixed content allowed
• no size restriction
• application knowledge

Main feature(s):

• Identifying single anomalous log lines
• AI/ML application, not rule based
• Taking context into account

Output(s):

• Anomalous line pointer
• In learning phase: feedback-form to teach what’s normal

• Service or development engineer, or annotated log files, needed for application-specific machine learning training
• Online access to logging computer, or local offline operation
• Processing power dependent on log file size and quantity

Machine and Instrument application developers, any service supplier who could exploit routinely logged information files.

Commercial licence to be negotiated.