Reference

Cognitive Analytics Data Model

The following diagram shows how the families used in Cognitive Analytics are related to one another.

Note: In the diagram, boxes represent entity families and arrows represent relationship families; the arrow points from the predecessor family to the successor family.

Cognitive Analytics facilitates the classification of both structured and unstructured data into a standardized, classified format. The Classified Equipment Standard family and the Classified Work History Standard family, represented in the image above, are used to store classified data when historical data from the Equipment family or the Work History family has been processed with a Cognitive Analytics cognition.

Families other than the Classified Equipment Standard family and the Classified Work History Standard family can be used to store classified data, but the relationship definition in the Has Classified Data relationship family must be modified to include them.

Alternatively, classified data can be stored in the existing Equipment and Work History families, but it should be stored in new, parallel fields to avoid overwriting the historical data.

Classified Equipment Standard Records

Classified Equipment Standard records store classified data produced by a Cognitive Analytics cognition that has processed historic equipment data. This topic provides an alphabetical list and description of the fields that exist for the Classified Equipment Standard family. The information in the table reflects the baseline state and behavior of these fields.

This family is enabled for site filtering, which means that records in this family can be assigned to a specific site, and will then only be accessible to users who are assigned to the same site and have the appropriate license and family privileges. For more information, refer to the Site Filtering section of the documentation.


Field	Data Type	Description	Behavior and Usage
Area	Character	The boundaries within a site or plant that describe where the equipment is located.	You can use this field to store classified data produced by a cognition.
Business	Character	If applicable, the name of the business within the company.	You can use this field to store classified data produced by a cognition.
Classification Cognition	Character	When this record was last processed, the name of the cognition that was executed.	You can use a Constant node to populate this field.
Classification Date	Date	When this record was last processed, the date on which the cognition was executed.	You can use the Now() function in a Constant node to populate this field.
Classification User	Character	When this record was last processed, the user who initiated the execution of the cognition.	You can use the User() function in a Constant node to populate this field.
Company	Character	The company name within the corporation.	You can use this field to store classified data produced by a cognition.
Corporation	Character	The corporation name.	You can use this field to store classified data produced by a cognition.
Corporation	Character	A field that indicates how important the piece of equipment is to production.	You can use this field to store classified data produced by a cognition.
Equipment Category	Character	The category of equipment, such as: Fixed Rotating Electrical Instruments	You can use this field to store classified data produced by a cognition.
Equipment Class	Character	An equipment class within the equipment category. For example, equipment classes under the Rotating equipment category might be Compressor, Pump, and Engine.	You can use this field to store classified data produced by a cognition.
Equipment ID	Character	A unique value that identifies each individual piece of equipment.	You can use this field to store classified data produced by a cognition.
Equipment Type	Character	The equipment type within the equipment class. For example, an equipment type under the Pump equipment class is Centrifugal.	You can use this field to store classified data produced by a cognition.
Functional Location ID	Character	The physical location of the equipment.	You can use this field to store classified data produced by a cognition.
Installation Date	Date	The installation date of the equipment.	You can use this field to store classified data produced by a cognition.
Latitude	Numeric	A numeric value representing the latitudinal position of the equipment.	You can use this field to store classified data produced by a cognition. Important: Values are stored as decimal degrees.
Longitude	Numeric	The numeric value representing the longitudinal position of the equipment.	You can use this field to store classified data produced by a cognition. Important: Values are stored as decimal degrees.
Manufacturer	Character	The name of the equipment manufacturer.	You can use this field to store classified data produced by a cognition.
Model Number	Character	The model number assigned to the equipment by the manufacturer.	You can use this field to store classified data produced by a cognition.
Region	Character	The geographic region in which the equipment is located.	You can use this field to store classified data produced by a cognition.
Short Description	Character	A brief description of the equipment.	You can use this field to store classified data produced by a cognition.
Site	Character	The name of the plant or site in which the equipment is located.	You can use this field to store classified data produced by a cognition.
Status	Character	The availability of the equipment.	You can use this field to store classified data produced by a cognition.
Technical ID	Character	The technical number of the equipment.	You can use this field to store classified data produced by a cognition.
Unit	Character	The process location of the equipment within an area in a site or plant.	You can use this field to store classified data produced by a cognition.
Unit Type	Character	The type of the material processed within a unit.	You can use this field to store classified data produced by a cognition.

Classified Work History Standard Records

Classified Work History Standard records store classified data produced by a Cognitive Analytics cognition that has processed historic work history data. This topic provides an alphabetical list and description of the fields that exist for the Classified Work History Standard family. The information in the table reflects the baseline state and behavior of these fields.

This family is enabled for site filtering, which means that records in this family can be assigned to a specific site, and will then only be accessible to users who are assigned to the same site and have the appropriate license and family privileges. For more information, refer to the Site Filtering section of the documentation.


Field	Data Type	Description	Behavior and Usage
Classification Cognition	Character	When this record was last processed, the name of the cognition that was executed.	You can use a Constant node to populate this field.
Classification Date	Date	When this record was last processed, the date on which the cognition was executed.	You can use the Now() function in a Constant node to populate this field.
Classification User	Character	When this record was last processed, the user who initiated the execution of the cognition.	You can use the User() function in a Constant node to populate this field.
Company	Character	The company name within the corporation.	You can use this field to store classified data produced by a cognition.
Corporation	Character	The corporation name.	You can use this field to store classified data produced by a cognition.
Detection Method	Character	A method used to detect potential events before they occur, which results in proactive work.	You can use this field to store classified data produced by a cognition.
Equipment ID	Character	The unique equipment ID for the piece of equipment that is associated with an event.	You can use this field to store classified data produced by a cognition.
Event Short Description	Character	A short description of the event.	You can use this field to store classified data produced by a cognition.
Event ID	Character	A unique work order number assigned to an event.	You can use this field to store classified data produced by a cognition.
Event Long Description	Text	A long description of the event.	You can use this field to store classified data produced by a cognition.
Event Start Date	Date	This date on which a work order is generated and assigned to a piece of equipment.	You can use this field to store classified data produced by a cognition.
Event Type	Character	The type of event, such as: Repair Preventative Maintenance (PM) Predictive Maintenance (PdM)	You can use this field to store classified data produced by a cognition.
Failure Mechanism	Character	The failure mechanism identified for a failure.	You can use the failureMechanism.py script within a Classifier node to populate this field.
Failure Mode	Character	The failure mode identified for a failure.	You can use this field to store classified data produced by a cognition.
Failure?	Boolean	A value that indicates whether or not a failure occurred. A failure occurs when a piece of equipment is unable to perform its required function.	You can use the isAFailure.py script within a Classifier node to populate this field.
Maintainable Item	Character	The maintainable item on the equipment involved in the event.	You can use the maintainableItem.py script within a Classifier node to populate this field.
Maintenance Completion Date	Date	The date on which maintenance work for a piece of equipment is finished.	You can use this field to store classified data produced by a cognition.
Maintenance Cost	Numeric	The total repair costs associated with a work event. This includes any materials, labor costs, and contractor costs.	You can use this field to store classified data produced by a cognition.
Maintenance Start Date	Date	The date that maintenance work for a piece of equipment begins.	You can use this field to store classified data produced by a cognition.
Observed Failure Mechanism	Character	A parallel field to store the original failure mechanism from the historic work history data after it has been translated into the corresponding standard value for failure mechanisms from the standard list.	You can use a Case node to translate your original values to the corresponding standard values. This translation will allow you to easily compare the original, observed failure mechanism against the prediction provided by the failureMechanism.py script in the Failure Mechanism field.
Observed Failure Mode	Character	A parallel field to store the original failure mode from the historic work history data after it has been translated into the corresponding standard value for failure modes from the standard list.	You can use this field to store classified data produced by a cognition.
Observed Failure?	Boolean	A parallel field to store the original failure value from the historic work history data after it has been translated into the corresponding standard value for failure from the standard list.	You can use a Case node to translate your original values to the corresponding standard values. This translation will allow you to easily compare the original, observed failure value against the prediction provided by the isAFailure.py script in the Failure? field.
Observed Maintainable Item	Character	A parallel field to store the original maintainable item from the historic work history data after it has been translated into the corresponding standard value for maintainable items from the standard list.	You can use a Case node to translate your original values to the corresponding standard values. This translation will allow you to easily compare the original, observed maintainable item against the prediction provided by the maintainableItem.py script in the Maintainable Item field.
Priority	Character	The importance of the work order compared to other work orders. The priority exists to help a planning department distribute resources and meet production requirements.	You can use this field to store classified data produced by a cognition.
Site	Character	The name of the plant or site in which the equipment associated with the work request is located.	You can use this field to store classified data produced by a cognition.
Status	Character	The status of the work event.	You can use this field to store classified data produced by a cognition.
Unit	Character	The process location of the equipment within a site or plant.	You can use this field to store classified data produced by a cognition.

Cognition Records

Cognition records store basic information about cognitions. This topic provides an alphabetical list and description of the fields that exist for the Cognition family and appear on the Cognition datasheet. The information in the table reflects the baseline state and behavior of these fields.

This family is not enabled for site filtering, which means that records in this family can be accessed by any user with the appropriate license and family privileges. For more information, refer to the Site Filtering section of the documentation.


Field	Data Type	Description	Behavior and Usage
Active?	Boolean	Indicates whether the cognition is active or inactive.	Reflects the selection of the activate/deactivate policy button in the Details workspace. When Active? is true, the cognition will execute at the scheduled interval. When Active? is false, the cognition will not execute.
Classification Relationship	Character	The relationship family that links the input data source to the output family.	This field does not appear in the Details workspace by default. If the family in the Output Family field is included in the relationship definition of the Has Classified Data relationship family, the Has Classified Data relationship family ID is populated automatically in this field. For all other families, this field remains empty.
Cognition Manual Start Date	Date	A date by which records from the input data can be filtered before cognition execution.	Contains the value that you specify in the Process Cognition data starting from this date box in the Details workspace. If a cognition is executed manually and the query supplying the input data has a query prompt on a date field, the value in this field is passed as a date parameter, and the query will only return records that satisfy the query prompt's logic. If a cognition is configured to be executed according to a schedule, on a successful execution, this field is populated automatically by the GE Digital APM system with the value in the Last Scheduled Execution field. See the About Cognition Execution topic for more information.
Configuration	Text	Configuration information for the cognition.	This field does not appear in the Details workspace by default. This field is populated automatically by the GE Digital APM System. Important: Do not modify the contents of this field.
Data Input	Character	The reference to the data source (e.g., query) to be processed by the cognition.	Contains the data source that you specify in the Bulk Input Data box in the Details workspace. The data structure found in the referenced data source provides the input fields to use when designing the cognition. Additionally, the data provided in the referenced data source is the data that will be processed when the cognition is executed. Note: The data source must include the Entity Key for the records. Ensure that the query you use is a Select query. Other types of queries will cause errors to occur.
Description	Text	A description of the cognition.	None
Last Scheduled Execution	Date	The end date for the scheduled execution of the cognition.	This field does not appear in the Details workspace by default. This field is populated automatically by the GE Digital APM System.
Layout	Text	The design layout of the cognition.	This field does not appear in the Details workspace by default. This field is populated automatically by the GE Digital APM System. Important: Do not modify the contents of this field.
Name	Character	The name of the cognition.	This field is required and must be unique.
Next Scheduled Execution	Date	The next date for a recurring scheduled execution of the cognition.	This field does not appear in the Details workspace by default. This field is populated automatically by the GE Digital APMM System.
Output Family	Character	The family to which the cognition will write classified data when executed.	Contains the name of the family that you select in the Classification Family box in the Details workspace. The data structure found in the referenced family provides the output fields to use when designing the cognition. Note: Only a family included in the relationship definition of the Has Classified Data relationship family and the same family that provides the data source in the Data Input field are available to be selected in this field.
Schedule	Character	Code defining the cognition's schedule.	This field does not appear in the Details workspace by default. This field is populated automatically by the GE Digital APM System.
Schedule Executed	Boolean	Indicates whether or not the currently configured schedule has executed at least once.	This field does not appear in the Details workspace by default. This field is populated automatically by the GE Digital APM System.
Schedule Time Zone	Character	The time zone associated with the scheduled execution of the cognition.	Contains the time zone that you specify in the Time Zone box in the Edit Schedule window.
Scheduled	Boolean	Indicates whether or not the cognition has a defined execution schedule.	This field does not appear in the Details workspace by default. This field is populated automatically by the GE Digital APM System.
Security Group	Numeric	Contains the security group that has privileges to modify the cognition.	This field contains the security group that you specify in the Security Group box in the Details workspace.

Dashboards Folder

The Catalog folder \\Public\Meridium\Modules\Cognitive Analytics\Dashboards contains the following items.


Dashboard Name	Behavior and Usage
Cognitive Analytics Overview	Displays the Cognitive Analytics Overview page.

Queries Folder

The Catalog folder \\Public\Meridium\Modules\Cognitive Analytics\Queries contains the following items.


Query Name	Behavior and Usage
Cognitive Overview - Classifiers	Returns the list of Machine Learning and Fuzzy Logic scripts (i.e., classifiers) that is displayed on the Cognitive Analytics Overview page.
Cognitive Overview - Cognitions	Returns the list of cognitions that is displayed on the Cognitive Analytics Overview page.

Scripts Folder

The Catalog folder \\Public\Meridium\Modules\Cognitive Analytics\Scripts contains the following items.


Script Name	Behavior and Usage
failureMechanism.py	The failureMechanism.py script is a Fuzzy Logic script that predicts the failure mechanism involved in a work history event. The failureMechanism.py script can be used in a Classifier node in a cognition so that historic, transactional data can be properly classified. Note: You should consider first identifying which work history events should be defined as failures before using failureMechanism.py. When used on all work history events, failureMechanism.py will provide a prediction regardless of the failure classification of the work history event. The failureMechanism.py script uses the following standard lists: failure_mechanism: provides valid values for classification predictions. brokenwords: provides a list of important words to identify during text cleaning. Due to constraints in legacy CMMS systems, words can appear to the script to be broken in the middle with a space. This list identifies potentially broken words that need to be modified and used. contraction: provides a list of contractions that are modified for consistency during text cleaning. stopwords: provides a list of words to remove during text cleaning. For best results when testing or using failureMechanism.py, create a query or dataset that includes as many of the following fields as possible: Work history event short description Work history event long description Failure mechanism description Failure mode description
Fuzzy Logic Template	The Fuzzy Logic Template script is the template Fuzzy Logic script that is used to create new Fuzzy Logic scripts. Important: Do not modify the contents of this script.
isAFailure.py	The isAFailure.py script is a Machine Learning script that predicts if a work history event represents a failure as defined by ISO 14224: 2006. The isAFailure.py script can be used in a Classifier node in a cognition so that historic, transactional data can be properly classified. The isAFailure.py script uses the following standard lists: failure: provides valid values for classification predictions. brokenwords: provides a list of important words to identify during text cleaning. Due to constraints in legacy CMMS systems, words can appear to the script to be broken in the middle with a space. This list identifies potentially broken words that need to be modified and used. contraction: provides a list of contractions that are modified for consistency during text cleaning. stopwords: provides a list of words to remove during text cleaning. The isAFailure.py script includes anonymized training data from GE Digital APM's Asset Answers database. The training data includes features from the following fields: Work history event short description Failure mode description Priority description When testing, using, or further training isAFailure.py, create a query or dataset that includes data with similar features.
Machine Learning Template	The Machine Learning Template script is the template Machine Learning script that is used to create new Machine Learning scripts. Important: Do not modify the contents of this script.
maintainableItem.py	The maintainableItem.py script is a Fuzzy Logic script that predicts the maintainable item involved in a work history event. The maintainableItem.py script can be used in a Classifier node in a cognition so that historic, transactional data can be properly classified The maintainableItem.py script uses the following standard lists: maintainable_item: provides valid values for classification predictions. brokenwords: provides a list of important words to identify during text cleaning. Due to constraints in legacy CMMS systems, words can appear to the script to be broken in the middle with a space. This list identifies potentially broken words that need to be modified and used. contraction: provides a list of contractions that are modified for consistency during text cleaning. stopwords: provides a list of words to remove during text cleaning. For best results when testing or using maintainableItem.py, create a query or dataset that includes as many of the following fields as possible: Work history event short description Work history event long description Maintainable item description

Cognitive Analytics URLs

There is one URL route associated with Cognitive Analytics: cognitive. The following table describes the various paths that build on the route, and the elements that you can specify for each.

Tip: For more information, refer to the URLs section of the documentation.


Element	Description	Accepted Value(s)	Notes
cognitive/overview: Displays the Cognitive Analytics Overview page.
cognitive/mapping/<EntityKey>/details: Opens a new or existing cognition in the Details workspace.
<EntityKey>	Specifies the cognition that you want to open.	Any numeric Entity Key that corresponds to an existing cognition.	Opens the specified cognition in the Details workspace. This value is required to open an existing cognition from a URL.
<EntityKey>	Specifies the cognition that you want to open.	0	Opens a new cognition in the Details workspace.

Example URLs


Example URL	Destination
cognitive/mapping/0/details	Details workspace for a new cognition.
cognitive/mapping/5082435/details	Details workspace for the cognition with the Entity Key 5082435.