Main
Overview
Getting Started
Components
Videos
Sample Files

simulation components, logic, intelligence, patterns
AIB2 Components - Pattern Recorder
Category: Decision Logic
  Description:
The Pattern Recorder component provides functionality for recording the combinations of values seen across all of the provided inputs connected to it.

Each cycle of execution computes the vector of feature values provided to it, comparing it against prior recordings. If this pattern has been seen, an increment of hits is made to the prior pattern. If the new pattern is unique it is recorded and given a unique ID for later retrieval and use.

Users can access the results of the recordings, however long in length, via the properties dialog box for the component. Within the dialog, all prior observations are provided for view in detail, along with observed number of occurrences. These patterns can be exported to CSV file by choosing Export and selecting a filename.
 

  Inputs: 1
Input Name:
Description:
Read Inputs (enable)
Boolean input of 1 or 0, ON or OFF, enabling or disabling pattern recording. If ON, a read of the input array is made, and a recording of the pattern is stored in a viewable file within the properties dialog box, or for export. Default is ON.

 

  Outputs: 6
Output Name:
Description:
Recorded Flag
Convenient output flag indicating that a recording just took place. Given the possibility that recordings may be configured in some systems to take place sporadically, this flag can be used by other components that take secondary actions when one occurs.
Assigned Pattern Number
This is the unique id of the new pattern recorded that has been assigned by the pattern recorder.
Unique Patterns Recorded
Total number of unique patterns that have been recorded since the last reset.
Total Patterns Recorded
Total number of patterns that have been recorded. This would include those seen that are in replication of prior observed patterns.
Patterns Possible
This is a calculated number simply available for convenience indicating how many possible unique patterns there are, given the number of features within the pattern.
Most Common Pattern
This is the ID of the pattern seen most frequently, since the last reset.

 

  Components in Same Category: Neural Network
Segmenter Single Out
Segmenter Multple Outs
Feedback Flow
Pattern Generator
Pattern Analyzer
Decision Matrix
Target Input Comparator
Pattern Sequencer
State Machine Simple
States Awareness Hub
State Machine Timer


  Sample Files with this Component:
 



Copyright © TinMan Systems, Inc. 2010-2015. All Rights Reserved.
Questions or comments: support@tinmansystems.com