Systems and methods for providing inquiry responses using linguistics and machine learning
A system is disclosed for automatically providing solutions to user questions containing text and data. The system includes a linguistic module to extract keywords from the question text, lookup solutions from a knowledge base that match the keywords, with a score based on the keyword frequency and match, and use a machine learning model trained on previous question data to predict solutions from the question data, with score based on the model's probability, and a combination module for combining the solutions ordered by their respective normalized scores.
1. A system for automatically providing solutions to user questions relating to an operation of an industrial machine, the system comprising:
a linguistic module that executes on a computer to extract keywords from a question text relating to an operation of an industrial machine, lookup a first plurality of solutions from a knowledge base that match said keywords, and assign each of the first plurality of solutions with a first score based on a number or a frequency of matching keywords in a text of the solution;
a machine learning module that executes on the computer to provide a predictive model trained on previous diagnostic data and corresponding solutions relating to the operation of other industrial machines, use the predictive model to predict a second plurality of solutions based on non-textual diagnostic data generated by the industrial machine, and assign each of the second plurality of solutions with a second score based on a model probability associated with the solution, wherein the non-textual diagnostic data represents at least a current state of the industrial machine, and
a combination module that executes on the computer to combine the first plurality of solutions and the second plurality of solutions in a ranked order determined by the first score and the second score of each solution.
2. The system as claimed in claim 1 , wherein the system further includes a translation module that executes on the computer for translating the question text into a common language.
3. The system as claimed in claim 1 , wherein the system further provides a linguistic analysis of the question text.
4. The system as claimed in claim 3 , wherein the linguistics analysis includes parsing the question text and classifying words in said text.
5. The system as claimed in claim 1 , wherein the machine learning module encodes the non-textual diagnostic data generated by the industrial machine into a numerical format suitable for prediction.
6. The system as claimed in claim 1 , wherein the machine learning module selectively excludes a portion of the diagnostic data having no correlation for predicting the second plurality of solutions.
7. The system as claimed in claim 1 , wherein the question text further includes an identity of the industrial machine, and wherein the computer is further configured to transmit a request for the non-textual diagnostic data to the industrial machine identified in the question text, receive the non-textual diagnostic data generated by the industrial machine in response to the request, and predict the second plurality of solutions using the machine learning model based on the non-textual diagnostic data received from the industrial machine.
8. The system as claimed in claim 1 , wherein the operation of an industrial machine includes at least one of a setup operation and a troubleshooting operation of the industrial machine.
9. A method for automatically providing solutions to user questions relating to a device, the method comprising:
extracting keywords from a question text relating to an operation of an industrial machine;
looking up a first plurality of solutions from a knowledge base that match said keywords;
assigning each of the first plurality of solutions with a first score based on a number or a frequency of matching keywords in a text of the solution;
obtaining non-textual diagnostic data generated by the industrial machine, wherein the non-textual diagnostic data represents at least a current state of the industrial machine;
predicting a second plurality of solutions using a machine learning model based on the non-textual diagnostic data generated by the industrial machine, the machine learning model being trained on previous diagnostic data and corresponding solutions relating to the operation of other industrial machines;
assigning each of a second plurality of solutions with a second score based on a model probability associated with the solution; and
combining the first plurality of solutions and the second plurality of solutions in a ranked order determined by the first score and the second score of each solution.
10. The method as claimed in claim 9 , further comprising translating the question text into a common language.
11. The method as claimed in claim 9 , further comprising providing a linguistic analysis of the question text.
12. The method as claimed in claim 11 , wherein providing the linguistics analysis includes parsing the question text and classifying words in said text.
13. The method as claimed in claim 9 , further comprising encoding the non-textual diagnostic data generated by the industrial machine into a numerical format suitable for prediction.
14. The method as claimed in claim 9 , further comprising selectively excluding a portion of the diagnostic data that has no correlation for predicting the second plurality of solutions.
15. The method as claimed in claim 9 , wherein the question text further includes an identity of the industrial machine, the method further comprising:
transmitting a request for the non-textual diagnostic data to the industrial machine identified in the question text;
receiving the non-textual diagnostic data generated by the industrial machine in response to the request; and
predicting the second plurality of solutions using the machine learning model based on the non-textual diagnostic data received from the industrial machine.
16. The method as claimed in claim 9 , wherein the operation of an industrial machine includes at least one of a setup operation and a troubleshooting operation of the industrial machine.