A batch profile creates a profile based on telemetry that was captured in the past. This is sometimes referred to as profile seeding or backfilling. Batch profiles can be used to understand the historical behaviors and trends of a profile to determine if the profile has predictive value for model building. You can use the streaming profiler and the batch profiler to gather and understand both current and historical behaviors and trends. This information can be used to determine if the profiler feature set matches reality and has predictive value for model building.