Method and system for verbal scale recognition using machine learning
A computer inputs data including different verbal judgment sets. Each different verbal judgment set includes words which are votes that define different rank values and each represents an evaluation of an alternative. The processor determines a word similarity score of each word in the verbal judgment sets to predefined words in a predefined scale. The processor determines a set similarity score between the different verbal judgment set and the predefined scale based on the words included in the different verbal judgment set and the predefined words within the predefined scale. The processor maps the words of the different verbal judgment sets to a numerical scale that corresponds to the predefined scale, based on the set similarity score. The processor interprets the different verbal judgment sets in the universe of known data based on the numerical scale and provides cleansed data which is used by a data-dependent application.
1. A computer system comprising:
at least one processor configured to:
input data including a plurality of different verbal judgment sets, wherein each one of the different verbal judgment sets includes words which are votes that define different rank values within the one of the different verbal judgment sets, wherein a rank value represents an evaluation of an alternative;
for each word of the words in the different verbal judgment sets:
determine a word similarity score of the word to predefined words in at least one predefined scale of a universe of predefined scales;
for the different verbal judgment sets:
determine a set similarity score between the one of the different verbal judgment sets and the at least one predefined scale based on the word similarity scores of the words included in the one of the different verbal judgment sets and the predefined words within the at least one predefined scale;
map the words of the one of the different verbal judgment sets to a numerical scale that corresponds to the at least one predefined scale of the universe of predefined scales, based on the set similarity score to the at least one predefined scale; and
interpret the different verbal judgment sets in a universe of known data based on the numerical scale and provide the interpreted different verbal judgment sets corresponding to the plurality of different verbal judgment sets.
2. The computer system of claim 1 , wherein the processor is further configured to
determine an implied rank value within the one of the different verbal judgment sets; and
adjust the numerical scale based on the one of the different verbal judgment sets including the implied rank value.
3. The computer system of claim 1 , wherein the processor is further configured to
receive additional data including the different verbal judgment sets,
determine a respective matching score between the at least one predefined scale and the different verbal judgment sets in the additional data,
when a perfect matching score is determined, update the numerical scale based on the predefined scale and the different verbal judgment sets with the perfect matching score.
4. The computer system of claim 1 , wherein the processor is further configured to
generate a confidence score that indicates how the different verbal judgment sets match the numerical scale.
5. The computer system of claim 1 , wherein the input data is based on actual votes in a decision system, wherein the processor is further configured to interpret the actual votes using the numerical scale.
6. The computer system of claim 1 , wherein the processor is further configured to infer a missing step as an additional rank value into the one of the different verbal judgment sets.
7. The computer system of claim 1 , wherein the processor is further configured to infer a missing concept as an additional rank value into the one of the different verbal judgment sets.
8. A method for recognizing a numerical scale as a precise numerical interpretation of a plurality of different verbal judgment sets, the method comprising:
inputting data including the plurality of different verbal judgment sets, wherein each one of the different verbal judgment sets includes words which are votes that define different rank values within the one of the different verbal judgment sets, wherein a rank value represents an evaluation of an alternative;
for each word of the words in the different verbal judgment sets:
determining a word similarity score of the word in the one of the different verbal judgment sets to predefined words in at least one predefined scale of a universe of predefined scales;
for the different verbal judgment sets:
determining a set similarity score between the one of the different verbal judgment sets and the at least one predefined scale based on the word similarity scores of the words included in the one of the different verbal judgment sets and the predefined words within the at least one predefined scale;
mapping the words of the one of the different verbal judgment sets to a numerical scale that corresponds to the at least one predefined scale of the universe of predefined scales, based on the set similarity score to the at least one predefined scale; and
interpreting the different verbal judgment sets in a universe of known data based on the numerical scale and provide the interpreted different verbal judgment sets corresponding to the plurality of different verbal judgment sets.
9. The method of claim 8 , further comprising
determining an implied rank value within the one of the different verbal judgment sets; and
adjusting the numerical scale based on the one of the different verbal judgment sets including the implied rank value.
10. The method of claim 8 , further comprising
receiving additional data including the different verbal judgment sets,
determining a respective matching score between the at least one predefined scale and the different verbal judgment sets in the additional data,
when a perfect matching score is determined, updating the numerical scale based on the predefined scale and the different verbal judgment sets with the perfect matching score.
11. The method of claim 8 , further comprising
generating a confidence score that indicates how the different verbal judgment sets match the numerical scale.
12. The method of claim 8 , wherein the input data is based on actual votes in a decision system, wherein the actual votes are interpreted using the numerical scale.
13. The method of claim 8 , further comprising inferring a missing step as an additional rank value into the one of the different verbal judgment sets.
14. The method of claim 8 , further comprising inferring a missing concept as an additional rank value into the one of the different verbal judgment sets.
15. A non-transitory computer-readable medium comprising instructions for execution by a computer, the instructions including a computer-implemented method for recognizing a numerical scale as a precise numerical interpretation of a plurality of different verbal judgment sets, the instructions for implementing:
inputting data including the plurality of different verbal judgment sets, wherein each one of the different verbal judgment sets includes words which are votes that define different rank values within the one of the different verbal judgment sets, wherein a rank value represents an evaluation of an alternative;
for each word of the words in the different verbal judgment sets:
determining a word similarity score of the word in the one of the different verbal judgment sets to predefined words in at least one predefined scale of a universe of predefined scales;
for the different verbal judgment sets:
determining a set similarity score between the one of the different verbal judgment sets and the at least one predefined scale based on the word similarity scores of the words included in the one of the different verbal judgment sets and the predefined words within the at least one predefined scale;
mapping the words of the one of the different verbal judgment sets to a numerical scale that corresponds to the at least one predefined scale of the universe of predefined scales, based on the set similarity score to the at least one predefined scale; and
interpreting the different verbal judgment sets in a universe of known data based on the numerical scale and providing the interpreted different verbal judgment sets corresponding to the plurality of different verbal judgment sets.
16. The non-transitory computer-readable medium of claim 15 , further comprising
determining an implied rank value within the one of the different verbal judgment sets; and
adjusting the numerical scale based on the one of the different verbal judgment sets including the implied rank value.
17. The non-transitory computer-readable medium of claim 15 , further comprising
receiving additional data including the different verbal judgment sets,
determining a respective matching score between the at least one predefined scale and the different verbal judgment sets in the additional data,
when a perfect matching score is determined, updating the numerical scale based on the predefined scale and the different verbal judgment sets with the perfect matching score.
18. The non-transitory computer-readable medium of claim 15 , further comprising
generating a confidence score that indicates how the different verbal judgment sets match the numerical scale.
19. The non-transitory computer-readable medium of claim 15 , wherein the input data is based on actual votes in a decision system, wherein the actual votes are interpreted using the numerical scale.
20. The non-transitory computer-readable medium of claim 15 , further comprising inferring a missing step as an additional rank value into the one of the different verbal judgment sets.
21. The non-transitory computer-readable medium of claim 15 , further comprising inferring a missing concept as an additional rank value into the one of the different verbal judgment sets.