Multidimensional data analysis for issue prediction

Patent metadata

Field	Value
Publication number	US-2022198259-A1
Application number	US-202017129367-A
Country	US
Kind code	A1
Filing date	Dec 21, 2020
Priority date	Dec 21, 2020
Publication date	Jun 23, 2022
Grant date	—

Abstract

A system for issue prediction based on multidimensional data analysis includes a model generator that receives a resolved data item relating to a service issue. The resolved data item includes different attributes corresponding to multiple data dimensions and adjusts a population of attributes based on a statistical data model and a deep learning data model operating independent of each other. The statistical data model operates on the attributes for providing a predictive feature and the deep learning data model operates on the attributes for providing a predictive label based on performance metrics related to the data dimensions. The predictive feature and the predictive label collectively define training data. The model generator also trains a classification model based on the training data for predicting a potential issue related to an unresolved data item. The trained data model provides a trigger based on the potential issue being related to the performance metrics.

First claim

I/We claim:

1 . A system comprising:
a processor; and a model generator coupled to the processor, the model generator to:
receive a resolved data item relating to a service issue, the resolved data item including a set of attributes corresponding to a plurality of predefined data dimensions, wherein the set of attributes relate to a historical resolution for the service issue, a category related to the service issue, a description of the service issue, an identity indicator, and a creation date with a first timestamp related to the resolved data item;
adjust a population of attributes in the set based on a plurality of data models including a statistical data model and a deep learning data model operating independent of each other, the statistical data model operating on the set for providing a predictive feature and the deep learning data model operating on the set for providing a predictive label based on predefined performance metrics related to the plurality of data dimensions, wherein the predictive feature and the predictive label collectively define training data; and
train a classification model based on the training data, the trained data model predicting a potential issue related to an unresolved data item, wherein the trained data model provides a trigger based on the potential issue being related to the predefined performance metrics used for obtaining the training data.

2 . The system of claim 1 , further comprising an action performer coupled to the processor, in response to the trigger, the action performer to:
classify the unresolved data item into a priority group related to the predefined performance metrics, wherein the priority group corresponds to one of an age, an escalation, and a sentiment; and perform a predefined action based on the priority group associated with the unresolved data item, the predefined action including at least one of configuring a data model for predicting a resolution for the potential issue, manipulating a position of the unresolved data item in a predefined queue, assigning a rank to the unresolved data item based on the priority group related thereto, communicating at least one of the potential issue and the unresolved data item to a predefined user or a predefined device, and initiating a predefined resolution associated with the priority group or the potential issue.

3 . The system of claim 1 , wherein the statistical data model and the deep learning data model operate simultaneously.

4 . The system of claim 1 , wherein the set of attributes further relate to at least one of a historical resolution date with a second timestamp and one of a geographical indicator, a satisfaction indicator, and a revenue indicator related to one of a user and an entity associated with the resolved data item.

5 . The system of claim 1 , wherein the resolved data item includes one of a service ticket, a message, and a query, or a combination thereof.

6 . The system of claim 1 , wherein the unresolved data item includes an attribute corresponding to the set of attributes associated with the resolved data item, and wherein the unresolved data item includes an outgoing message created in response to the potential issue.

7 . The system of claim 1 , wherein the plurality of predefined data dimensions includes temporal data, service data, natural language text data, and performance data, or a combination thereof, the service data including at least one of a service process, a service category, a service sub-category, a service issue category, and a service issue sub-category and the performance data including at least one of a volume of data items corresponding to the service issue, revenue data, and year-to-date satisfaction data related to one of a user and an entity, wherein each of the service data and the performance data corresponds to target values as per the respective predefined performance metrics.

8 . A method comprising:
receiving, by a processor, a resolved data item relating to a service issue, the resolved data item including a set of attributes corresponding to a plurality of predefined data dimensions, wherein the set of attributes relate to a historical resolution for the service issue, a category related to the service issue, a description of the service issue, an identity indicator, and a creation date with a first timestamp related to the resolved data item; adjusting, by the processor, a population of attributes in the set based on a plurality of data models including a statistical data model and a deep learning data model operating independent of each other, the statistical data model operating on the set for providing a predictive feature and the deep learning data model operating on the set for providing a predictive label based on predefined performance metrics related to the plurality of data dimensions, wherein the predictive feature and the predictive label collectively define training data; and training, by the processor, a classification model based on the training data, the trained data model predicting a potential issue related to an unresolved data item, wherein the trained data model provides a trigger based on the potential issue being related to the predefined performance metrics used for obtaining the training data.

9 . The method of claim 8 , further comprising in response to the trigger:
classifying, by the processor, the unresolved data item into a priority group related to the predefined performance metrics, wherein the priority group corresponds to one of an age, an escalation, and a sentiment; and performing, by the processor, a predefined action based on the priority group associated with the unresolved data item, the predefined action including at least one of configuring a data model for predicting a resolution for the potential issue, manipulating a position of the unresolved data item in a predefined queue, assigning a rank to the unresolved data item based on the priority group related thereto, communicating at least one of the potential issue and the unresolved data item to a predefined user or a predefined device, and initiating a predefined resolution associated with the priority group or the potential issue.

10 . The method of claim 8 , wherein the statistical data model and the deep learning data model operate simultaneously.

11 . The method of claim 8 , wherein the set of attributes further relate to at least one of a historical resolution date with a second timestamp and one of a geographical indicator, a satisfaction indicator, and a revenue indicator related to one of a user and an entity associated with the resolved data item.

12 . The method of claim 8 , wherein the resolved data item includes one of a service ticket, a message, and a query, or a combination thereof.

13 . The method of claim 8 , wherein the unresolved data item includes an attribute corresponding to the set of attributes associated with the resolved data item, and wherein the unresolved data item includes an outgoing message created in response to the potential issue.

14 . The method of claim 8 , wherein the plurality of predefined data dimensions includes temporal data, service data, natural language text data, and performance data, or a combination thereof, the service data including at least one of a service process, a service category, a service sub-category, a service issue category, and a service issue sub-category and the performance data including at least one of a volume of data items corresponding to the service issue, revenue data, and year-to-date satisfaction data related to one of a user and an entity, wherein each of the service data and the perfo

Assignees

• Accenture Global Solutions Ltd

Inventors

• Dutt Anindya
• Kuduva Kamalesh Kuppusamy
• Ramesh Prashanth
• Swamy Siddesha
• Khan Mohd Israil
• Narain Ankur
• Viswanathan Kumar

Classifications

• G06N3/045

  Combinations of networks · CPC title

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06N3/0985

  Hyperparameter optimisation; Meta-learning; Learning-to-learn · CPC title

• G06N3/0455

  Auto-encoder networks; Encoder-decoder networks · CPC title

• G06N3/09

  Supervised learning · CPC title