Operational business intelligence measurement and learning system

• US 10,467,550 B1
• Filed: 04/29/2017
• Issued: 11/05/2019
• Est. Priority Date: 07/12/2013
• Status: Active Grant

• Current Assignee: OpsVeda, Inc.
• Original Assignee: OpsVeda, Inc.
• Inventors: Gupta, Sanjiv, Bhoot, Abhishek, Somani, Dinesh
• Primary Examiner(s): Brown, Sheree N
• Application Number:
  US15/582,643
• Time in Patent Office: 920 Days
• Field of Search: 707600
• CPC Class Codes: G06F 16/22 Indexing; Data structures therefor; Storage structures
• CPC Class Codes: G06F 7/023 adaptive, e.g. self learning
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0633 Lists, e.g. purchase orders, compilation or processing

Abstract

An automated method of detecting patterns corresponding to a plurality of real world business measures corresponding to a plurality of business processes, assessing the next instance of such measures and related business attributes, and describing the next best action to optimize business outcomes based upon a plurality of control parameters. The system operates by continuously abstracting input data from a process agnostic data system (PADS) that links real-world things, activities and processes, into a process agnostic measure store (PAMS) configured to accept measures data without limitation as to a specific process or a plurality of processes. The machine self-learning system can then automatically project a business outcome, suggest most relevant attributes that can impact the said outcome, and suggest actions to change such outcome(s).

Claims

1. An automated method of executing machine self-learning methods to detect and report measurements of a plurality of real-world processes, said real world processes comprising a plurality of real-world things and/or real-world activities, and wherein at least some of said real-world things and/or activities are linked together to form real-world processes, said method comprising:

receiving, by at least one computer processor, raw data properties pertaining to said real-world things and activities, said raw data properties comprising when-where data, attribute data, identifier data, and quantity data associated with said real-world things and activities;

wherein at least some attribute data comprise link attribute data, and wherein at least some identifier data comprise thing-activity identifier data;

creating and storing, by said at least one computer processor, in real time as said raw data properties are received, said raw data properties, thing-activity identifier data and any link attribute data in a process aganostic database system (PADS) database;

as PADS database objects, said PADS database objects comprising object header information and at least one set of when-where data, attribute data, thing-activity identifier data, quantity data and object exception information;

continually executing measures rules to select data from said PADS database objects, irrespective of process, according to any of date, quantity/amount, and location based parameters, and continually operating on said measures rules selected data with measures rules selected transformations, thus continually creating measures rules selected and transformed data, and continually storing said measures rules selected and transformed data in a process agnostic measure store (PAMS) database;

wherein said measurements rules comprise algorithms for producing time-adjusted and corrected data obtained from said PADS database; and

wherein said transformations comprise algorithms to fit at least some of said measurements into at least one model showing trends in said measurements over time;

and executing machine self-leaning methods to analyze said PAMS database and output a plurality of measurements according to said measures rules and measures transformations.

2. The method of claim 1, wherein said machine self-learning methods are implemented on a machine self-learning system comprising an Experience Knowledge Database (EKDB).

3. The method of claim 1, wherein at least some of said raw data properties are received from a human or automated user by a remote computerized device and a network connection;

wherein said remote computerized device has at least one remote processor;

wherein said at least one remote processor is configured to obtain said raw data properties from said user, and automatically link said raw data properties to said object header information;

thereby receiving, by said at least one remote processor and said network connection, raw data properties from said user along with said object header information.

4. The method of claim 1, wherein at least some of said raw data properties are received from a human or automated user by a remote computerized device and a network connection;

wherein said remote computerized device has at least one remote processor;

wherein said at least one remote processor is configured to obtain said raw data properties from said user, and automatically link said raw data properties to said thing-activity identifier data;

thereby receiving, by said at least one remote processor and said network connection, raw data properties from said user along with said thing-activity identifier data.

5. The method of claim 1, wherein said real-world processes are business processes, said real world things comprise products that are goods, and wherein said activities comprise activities selected from the group consisting of putting products on hold, releasing products for delivery, shipping products, product sales, product returns, changing one or more raw materials to work-in-process goods or finished goods, creating purchase orders for products, creating purchase orders for raw materials for a product, delay in receiving products, and delay in receiving raw materials for a product;

or

wherein said real world things are selected from the group consisting of customers, vendors, raw materials, materials, and finished goods

or wherein said real world activities comprise various material movement &

manufacturing steps of a manufacturing &

distribution process, marketing activities, financial activities, real estate activities, personnel activities, regulatory activities, and quality assurance activities.

6. The method of claim 1, wherein

said PADS database objects comprise process agnostic objects that handle said raw data according to generic data models and machine self-learning processor imlemented methods executed on said at least one computer processor;

wherein said object header information is received along with said raw data properties, or wherein said object header information is obtained by said at least one computer processor to compare said raw data properties against a previously defined set of categories of real-world things and activities, and classifying said raw data according to said previously defined categories of real-world things and activities, thereby determining said object header information;

wherein said object exception information is obtained by said at least one computer processor to compare said raw data properties against a previously defined baseline properties of said previously defined set of real-world things and activities to further determine if any of said raw data properties represent an exception from said baseline properties, and if said exception is found, storing it as object exception information;

wherein PADS database objects comprising at least one stored object exception information are exception marked PADS database objects;

linking, by said at least one computer processor, at least some different PADS database objects comprising data pertaining to real-world things and activities together to form real-world processes by setting said link attribute data in said different PADS database objects to create said links according to at least one set of process linking rules;

wherein PADS database objects linked together by setting said linked attribute data are linked PADS database objects; and

wherein said method continually uses said measures rules to select data from said PADS database objects.

7. The method of claim 6, wherein analyzing, by said at least one computer processor and at least one algorithm, said exception marked PADS database objects and linked PADS database objects and determine an impact of said exception marked PADS database objects and linked PADS database objects on a status of at least some of said real world processes, and updating said PAMS database, is done either on a real-time basis as said data properties are received, or on a non-real time basis as determined by schedules set according to other criteria.

8. The method of claim 6, wherein a given PADS database object comprising data pertaining to real world activities is linked with zero or more different PADS database objects comprising data pertaining to real world things, and wherein said given PADS database object is also linked with zero or more different PADS database objects comprising data pertaining to different real world activities;

and

further altering, by said at least one processor, at least some of said links according to at least one set of business process link configuration rules.

9. The method of claim 6, further analyzing, by said at least one computer processor and at least one algorithm, said exception marked PADS database objects and linked PADS database objects and determining an impact of said exception marked PADS database objects and linked PADS database objects on a status of at least some of said real world processes, thereby determining real-world process exceptions;

and outputting, by said at least one computer processor, data pertaining to at least some of said real world process exceptions.

10. The method of claim 9, further analyzing, by said at least one computer processor, said machine self-learning methods, and impact analysis rules, said real world process exceptions according to said object exception information, and said link attribute data, of their respective PADS database objects, thereby determining at least one of an urgency, magnitude of impact, and cost of inaction the impact of said exceptions on said real world processes, and outputting data pertaining to at least some of said real world process exceptions according to at least one of said urgency, magnitude of impact, and cost of inaction of said exceptions on said real world processes.

11. An automated method of executing machine self-learning methods to detect and report measurements of a plurality of real-world processes, said real world processes comprising a plurality of real-world things and/or real-world activities, and wherein at least some of said real-world things and/or activities are linked together to form real-world processes, said method comprising:

receiving, by at least one computer processor, raw data properties pertaining to said real-world things and activities, said raw data properties comprising when-where data, attribute data, identifier data, and quantity data associated with said real-world things and activities;

wherein at least some attribute data comprise link attribute data, and wherein at least some identifier data comprise thing-activity identifier data;

creating and storing, by said at least one computer processor, in real time as said raw data properties are received, said raw data properties, thing-activity identifier data and any link attribute data in a process agnostic database system (PADS) database as PADS database objects, said PADS database objects comprising object header information and at least one set of when-where data, attribute data, thing-activity identifier data, quantity data and object exception information;

said PADS database objects being process agnostic objects that handle said raw data according to generic data models and machine self-learning methods executed on said at least one computer processor;

wherein said object header information is received along with said raw data properties, or wherein said object header information is obtained by said at least one computer processor to compare said raw data properties against a previously defined set of categories of real-world things and activities, and classifying said raw data according to said previously defined categories of real-world things and activities, thereby determining said object header information;

wherein said object exception information is obtained by said at least one computer processor to compare said raw data properties against a previously defined baseline properties of said previously defined set of real-world things and activities to further determine if any of said raw data properties represent an exception from said baseline properties, and if said exception is found, storing it as object exception information;

wherein PADS database objects comprising at least one stored object exception information are exception marked PADS database objects;

linking, by said at least one computer processor, at least some different PADS database objects comprising data pertaining to real-world things and activities together to form real-world processes by setting said link attribute data in said different PADS database objects to create said links according to at least one set of process linking rules;

wherein PADS database objects linked together by setting said linked attribute data are linked PADS database objects;

b) continually executing measures rules to select data from said PADS database objects, irrespective of process, according to any of date, quantity/amount, and location based parameters, and continually operating on said measures rules selected data with measures rules selected transformations, thus continually creating measures rules selected and transformed data, and continually storing said measures rules selected and transformed data in a process agnostic measure store (PAMS) database;

and executing machine self-leaning methods to analyze said PAMS database and output a plurality of measurements according to said measures rules and measures transformations.

12. The method of claim 11, wherein said measurements rules comprise algorithms for producing time-adjusted and corrected data obtained from said database;

and

wherein said transformations comprise algorithms to fit at least some of said measurements into at least one model showing trends in said measurements over time.

13. The method of claim 11, wherein said machine self-learning methods are implemented on a machine self-learning system comprising an Experience Knowledge Database (EKDB).

14. The method of claim 11, further analyzing, by said at least one computer processor and at least one algorithm, said exception marked PADS database objects and linked PADS database objects and determining an impact of said exception marked PADS database objects and linked PADS database objects on a status of at least some of said real world processes, thereby determining real-world process exceptions;

and outputting, by said at least one computer processor, data pertaining to at least some of said real world process exceptions.

15. A machine self-learning system for automatically detecting and reporting measurements of a plurality of real-world processes, said real world processes comprising a plurality of real-world things and/or real-world activities, and wherein at least some of said real-world things and/or activities are linked together to form real-world processes, said system comprising:

at least one computer processor and memory;

process agnostic database system (PADS) database object management methods stored in said memory, said PADS database object management methods configured to direct said at least one computer processor to receive raw data properties pertaining to said real-world things and activities, said raw data properties comprising when-where data, attribute data, identifier data, and quantity data associated with said real-world things and activities;

wherein at least some attribute data comprise link attribute data, and wherein at least some identifier data comprise thing-activity identifier data;

said PADS database object management methods further configured to direct said at least one computer processor to create and store in said memory, in real-time as said raw data properties are received, said raw data properties, thing-activity identifier data and any link attribute data in a PADS database as PADS database objects, said PADS database objects comprising object header information and at least one set of when-where data, attribute data, thing-activity identifier data, quantity data and object exception information;

said PADS database objects being process agnostic objects that handle said raw data, according to generic data models and at least one computer processor configured to run machine self-learning methods;

wherein said PADS database object management methods are further configured to direct said at least one computer processor to receive said object header information along with said raw data properties, and/or wherein said PADS database object management methods are further configured to direct said at least one computer processor to obtain said object header information by comparing said raw data properties against a previously defined set of categories of real-world things and activities, and to classify said raw data according to said previously defined categories of real-world things and activities, thereby determining said object header information;

wherein said PADS database object management methods are further configured to direct said at least one computer processor to link at least some different PADS database objects comprising data pertaining to real-world things and activities together to form real-world processes by setting said link attribute data in said different PADS database objects to create said links according to at least one set of process linking rules;

wherein PADS database objects linked together by setting said linked attribute data are linked PADS database objects;

wherein said at least one computer processor is further configured to measures rules to select data from said PADS database objects, irrespective of process, according to any of date, quantity/amount, and location based parameters, and continually operate on said measures rules selected data with measures rules selected transformations, thus continually creating measures rules selected and transformed data, and to continually store said measures rules selected and transformed data in a process agnostic measure store database (PAMS) database; and

wherein said at least one computer processor is further configured to use machine self-leaning methods to analyze said PAMS database and output a plurality of measurements according to said measures rules and measures transformations.

16. The system of claim 15, wherein said measurements rules comprise algorithms for producing time-adjusted and corrected data obtained from said database;

and

wherein said transformations comprise algorithms to fit at least some of said measurements into at least one model showing trends in said measurements over time.

17. The system of claim 15, wherein said machine self-learning methods further implement a machine self-learning system comprising an Experience Knowledge Database (EKDB).

18. The system of claim 15, wherein said PADS database object management methods are further configured to direct said at least one computer processor to obtain said object exception information by comparing said raw data properties against a previously defined baseline properties of said previously defined set of real-world things and activities to further determine if any of said raw data properties represent an exception from said baseline properties, and if said exception is found, storing it as object exception information;

and

wherein PADS database objects comprising at least one stored object exception information are exception marked PADS database objects.

19. The system of claim 18, wherein said PADS database object management methods are further configured to direct said at least one computer processor configured to execute at least one algorithm to analyze said exception marked PADS database objects and linked PADS database objects and determine an impact of said exception marked PADS database objects and linked PADS database objects on a status of at least some of said real world processes, thereby determining real-world process exceptions;

and wherein said PADS database object management methods are further configured to direct said at least one computer processor to output data pertaining to at least some of said real world process exceptions.