Understanding user sentiment using implicit user feedback in adaptive dialog systems
Various embodiments are provided for understanding user sentiment in a dialog system in a computing environment by a processor. A sentiment of a user may be detected according to a sentiment analysis and user feedback during a dialog with the user. One or more reasons for the sentiment of the user may be identified. Behavior of the dialog system may be adjusted according to the one or more reasons.
1. A method, by a processor, understanding user sentiment in a dialog system in a dialog system in a computing environment comprising:
detecting a sentiment of a user according to a sentiment analysis and user feedback during a dialog with the user;
identifying one or more reasons for the sentiment of the user according to a detected change in the sentiment of the user identified subsequent to the dialog system performing an action requested by the user during the dialog, wherein identifying the one or more reasons for the detected change in the sentiment includes determining that the action requested by the user was deficiently performed by the dialog system;
adjusting behavior of the dialog system according to the one or more reasons to correct the deficient action and restore the sentiment of the user to a state previous to the detected change; and
commensurate with the detecting, identifying, and adjusting, initiating a machine learning mechanism to perform one or more machine learning operations to collect the user feedback, perform a semantic analysis, train a classifier, learn contextual data associated with the dialog, and implement one or more corrective actions to adjust the behavior of the dialog system to correct the deficient action.
2. The method of claim 1 , further including inferring the sentiment of the user according to a voice analysis.
3. The method of claim 1 , further including inferring the sentiment of the user according to a video analysis.
4. The method of claim 1 , further including classifying the sentiment as a negative sentiment or a positive sentiment.
5. The method of claim 1 , further including assigning a sentiment score to each utterance of the user.
6. The method of claim 5 , further including classifying the sentiment according to the sentiment score, wherein the sentiment score indicates a degree of a negative sentiment or a degree of a positive sentiment.
7. A system, for understanding user sentiment in a dialog system in a computing environment, comprising:
one or more processors with executable instructions that when executed cause the system to:
detect a sentiment of a user according to a sentiment analysis and user feedback during a dialog with the user;
identify one or more reasons for the sentiment of the user according to a detected change in the sentiment of the user identified subsequent to the dialog system performing an action requested by the user during the dialog, wherein identifying the one or more reasons for the detected change in the sentiment includes determining that the action requested by the user was deficiently performed by the dialog system;
adjust behavior of the dialog system according to the one or more reasons to correct the deficient action and restore the sentiment of the user to a state previous to the detected change; and
commensurate with the detecting, identifying, and adjusting, initiate a machine learning mechanism to perform one or more machine learning operations to collect the user feedback, perform a semantic analysis, train a classifier, learn contextual data associated with the dialog, and implement one or more corrective actions to adjust the behavior of the dialog system to correct the deficient action.
8. The system of claim 7 , wherein the executable instructions further infer the sentiment of the user according to a voice analysis.
9. The system of claim 7 , wherein the executable instructions further infer the sentiment of the user according to a video analysis.
10. The system of claim 7 , wherein the executable instructions further classify the sentiment as a negative sentiment or a positive sentiment.
11. The system of claim 7 , wherein the executable instructions further assign a sentiment score to each utterance of the user.
12. The system of claim 11 , wherein the executable instructions further classify the sentiment according to the sentiment score, wherein the sentiment score indicates a degree of a negative sentiment or a degree of a positive sentiment.
13. A computer program product for, by one or more processors, understanding user sentiment in a dialog system in a computing environment, the computer program product comprising a non-transitory computer-readable storage medium having computer-readable program code portions stored therein, the computer-readable program code portions comprising:
an executable portion that detects a sentiment of a user according to a sentiment analysis and user feedback during a dialog with the user;
an executable portion that detects identifies one or more reasons for the sentiment of the user according to a detected change in the sentiment of the user identified subsequent to the dialog system performing an action requested by the user during the dialog, wherein identifying the one or more reasons for the detected change in the sentiment includes determining that the action requested by the user was deficiently performed by the dialog system;
an executable portion that adjusts behavior of the dialog system according to the one or more reasons to correct the deficient action and restore the sentiment of the user to a state previous to the detected change; and
an executable portion that, commensurate with the detecting, identifying, and adjusting, initiates a machine learning mechanism to perform one or more machine learning operations to collect the user feedback, perform a semantic analysis, train a classifier, learn contextual data associated with the dialog, and implement one or more corrective actions to adjust the behavior of the dialog system to correct the deficient action.
14. The computer program product of claim 13 , further including an executable portion that infers the sentiment of the user according to a voice analysis.
15. The computer program product of claim 13 , further including an executable portion that infers the sentiment of the user according to a video analysis.
16. The computer program product of claim 13 , further including an executable portion that classifies the sentiment as a negative sentiment or a positive sentiment.
17. The computer program product of claim 13 , further including an executable portion that:
assigns a sentiment score to each utterance of the user; and
classifies the sentiment according to the sentiment score, wherein the sentiment score indicates a degree of a negative sentiment or a degree of a positive sentiment.