Recommendation system for change management in a quality management system

1. A change management system for Quality Management System (QMS) dependency visualization, comprising:

a server, coupled to a processor, and configured to execute instructions that:

receive an input from a user, the input comprising information relating to a proposed change to one or more objects (QMS objects) in a QMS (QMS network) via a graphical user interface (GUI), comprising a written description and justification for the change.

apply natural language processing techniques to extract structured metadata from the one or more QMS objects; and

generate an impact score corresponding to aggregate risk associated with a change in a first QMS object towards all other QMS objects in the QMS, wherein risk is defined as a statistical measure of the likelihood that a change in one or more interdependent QMS objects may contradict or invalidate the one or more policies of another of the QMS objects,

wherein the impact of a change may be determined in a bi-directional manner, such that for any pair of objects (a,b) the risks of a directed impact (a impacting b, a->b, or b impacting a, b->a) are potentially asymmetrical |R(a->b)-R(b->a)|>=0,

wherein the QMS network is an intranet, an extranet, a virtual private network (VPN), a non-TCP/IP based network, or any LAN,

wherein the metadata features comprise latent references to other objects within the QMS, either by name or unique identifier, and

wherein the metadata features comprising one or more keywords and/or vectorized topic representations.

2. The change management system of claim 1 , the server further configured to execute instructions that:

aggregate QMS objects to form a directed graph data structure;

generate a directed graph representation of a dependency network corresponding to the QMS based on the aggregated QMS objects, the directed graph data structure being formed to link all pairs of objects in the QMS having a dependency relation implied by their respective metadata features and/or an inter-object linkage informed directly by user input;

store the directed graph representation in a searchable computer database.

3. The change management system of claim 1 , the server further configured to execute instructions that:

alert the user to the risk according to the statistical attributes of the derived dependency network; and

provide one or more recommendations for actions identify or mitigate the risk.

4. The change management system of claim 1 , wherein the impact score of a specific object that may vary over time even if that object itself has not changed, wherein the impact potential of an object is almost entirely dependent on the topology of the entire dependency network,

wherein non-localized changes to one or many objects may significantly impact risk calculations for all other members of the QMS, and

wherein the topology of the dependency network will also vary as a function of the addition and deletion of new obj ects over time.

5. The change management system of claim 1 , the server further configured to execute instructions that alert the user to changes in the impact score of one or many objects and generate visual representations in the GUI reflecting the temporal variation of such statistics over time.

6. The change management system of claim 1 , the server further configured to execute instructions that:

receive a query of one or more QMS objects input from the user;

perform a query of the QMS database according to the metadata features extracted from the input QMS objects and based on the connectivity of the input QMS objects to other objects implied by the metadata features;

determine relevance of objects to the query based on the combination of search parameters and the proximity of matching objects in the dependency network; and

determine search results associated with the impact score based on a parametric network search capability.

7. The change management system of claim 1 , the server further configured to execute instructions that include:

generate, in the GUI, an interactive visualization representation of the QMS objects, the dependency network, and to one or more inter-object relationship attributes, wherein the one or more inter-object relationship attributes comprise a risk score.

8. A computer-implemented method, comprising:

receiving an input from a user, the input comprising information relating to a proposed change to one or more objects (QMS objects) in a QMS (QMS network) via a graphical user interface (GUI), comprising a written description and justification for the change;

applying natural language processing techniques to extract structured metadata from the QMS objects;

aggregating the QMS objects to form a directed graph data structure;

generating a directed graph representation of a dependency network corresponding to the QMS based on the aggregated QMS objects, the directed graph data structure being formed to link all pairs of objects in the QMS having a dependency relation implied by their respective metadata features and/or an inter-object linkage informed directly by user input;

storing the directed graph representation in a searchable computer database; and

generating an impact score corresponding to aggregate risk associated with a change in a first QMS object towards all other QMS objects in the QMS, wherein risk is defined as a statistical measure of the likelihood that a change in one or more interdependent QMS objects may contradict or invalidate the one or more policies of another of the QMS objects, and

wherein the impact of a change may be determined in a bi-directional manner, such that for any pair of objects (a,b) the risks of a directed impact (a impacting b, a->b, or b impacting a, b->a) are potentially asymmetrical |R(a->b)-R(b->a)|>=0,

wherein the QMS network is an intranet, an extranet, a virtual private network (VPN), a non-TCP/IP based network, or any LAN,

wherein the metadata features comprise latent references to other objects within the QMS, either by name or unique identifier, and

wherein the metadata features comprising one or more keywords and/or vectorized topic representations.

9. The computer-implemented method of claim 8 , further comprising:

alerting the user to the risk according to the statistical attributes of the derived dependency network; and

providing one or more recommendations for actions identify or mitigate risk.

10. The computer-implemented method of claim 8 , wherein an impact score of the first QMS object may vary over time, even if the first QMS object has not changed, wherein an impact potential of an object is based on the topology of the entire dependency network,

wherein non-localized changes to one or many objects may significantly impact risk calculations for all other members of the QMS, and

wherein the topology of the dependency network can vary as a function of the addition and deletion of new objects over time.

11. The computer-implemented method of claim 8 , further comprising:

alerting the user to changes in the impact score of one or more QMS objects; and

generating visual representations in the GUI reflecting the temporal variation of such statistics over time.

12. The computer-implemented method of claim 8 , further comprising:

receiving a query of one or more QMS objects input from the user;

performing a query of the QMS database according to the metadata features extracted from the input QMS objects and based on the connectivity of the input QMS objects to other objects implied by the metadata features;

determining relevance of objects to the query based on the combination of search parameters and the proximity of matching objects in the dependency network; and

determining search results associated with the impact score based on a parametric network search capability.

13. A non-transitory tangible computer-readable device having instructions stored thereon that, when executed by a computing device, cause the computing device to perform operations comprising:

receiving an input from a user, the input comprising information relating to a proposed change to one or more objects (QMS objects) in a QMS (QMS network) via a graphical user interface, comprising a written description and justification for the change;

applying natural language processing techniques to extract structured metadata from the QMS objects;

aggregating the QMS objects to form a directed graph data structure;

generating a directed graph representation of a dependency network corresponding to the QMS based on the aggregated QMS objects, the directed graph data structure being formed to link all pairs of objects in the QMS having a dependency relation implied by their respective metadata features and/or an inter-object linkage informed directly by user input;

storing the directed graph representation in a searchable computer database; and

generating an impact score corresponding to aggregate risk associated with a change in a first QMS object towards all other QMS objects in the QMS, wherein risk is defined as a statistical measure of the likelihood that a change in one or more interdependent QMS objects may contradict or invalidate the one or more policies of another of the QMS objects, and

wherein the impact of a change may be determined in a bi-directional manner, such that for any pair of objects (a,b) the risks of a directed impact (a impacting b, a->b, or b impacting a, b->a) are potentially asymmetrical |R(a->b)-R(b->a)|>=0,

wherein the QMS network is an intranet, an extranet, a virtual private network (VPN), a non-TCP/IP based network, or any LAN,

wherein the metadata features comprise latent references to other objects within the QMS, either by name or unique identifier, and

wherein the metadata features comprising one or more keywords and/or vectorized topic representations.

14. The computer-readable device of claim 13 , the operations further comprising:

alerting the user to the risk according to the statistical attributes of the derived dependency network; and

providing one or more recommendations for actions identify or mitigate risk.

15. The computer-readable device of claim 13 , wherein an impact score of the first QMS object may vary over time, even if the first QMS object has not changed,

wherein an impact potential of an object is based on the topology of the entire dependency network,

wherein non-localized changes to one or many objects may significantly impact risk calculations for all other members of the QMS, and

wherein the topology of the dependency network can vary as a function of the addition and deletion of new objects over time.

16. The computer-readable device of claim 13 , the operations further comprising:

alerting the user to changes in the impact score of one or more QMS objects; and

generating visual representations reflecting the temporal variation of such statistics over time.

17. The computer-readable device of claim 13 , the operations further comprising:

receiving a query of one or more QMS objects input from the user;

performing a query of the QMS database according to the metadata features extracted from the input QMS objects and based on the connectivity of the input QMS objects to other objects implied by the metadata features;

determining relevance of objects to the query based on the combination of search parameters and the proximity of matching objects in the dependency network; and

determining search results associated with the impact score based on a parametric network search capability.