Identification of attributes and values using multiple classifiers
A body of text comprises a plurality of unknown attributes and a plurality of unknown values. A first classification sub-component labels a first portion of the plurality of unknown values as a first set of values, whereas a second classification sub-component labels a portion of the plurality of unknown attributes as a set of attributes and a second portion of the plurality of unknown values as a second set of values. Learning models implemented by the first and second classification subcomponents are updated based on the set of attributes and the first and second set of values. The first classification sub-component implements at least one supervised classification technique, whereas the second classification sub-component implements an unsupervised and/or semi-supervised classification technique. Active learning may be employed to provide at least one of a corrected attribute and/or corrected value that may be used to update the learning models.
1. A method for at least one processing device to identify at least one attribute and at least one value in a body of text comprising a plurality of unknown attributes and a plurality of unknown values, the method comprising:
labeling, by a first classification sub-component operating on a first portion of the body of text and implemented by the at least one processing device, a first portion of the plurality of unknown values as a first set of values;
labeling, by a second classification sub-component operating on a second portion of the body of text and implemented by the at least one processing device, a portion of the plurality of unknown attributes as a set of attributes and a second portion of the plurality of unknown values as a second set of values; and
updating, by the first classification sub-component and the second classification sub-component, learning models implemented by the first classification sub-component and the second classification sub-component based on the set of attributes and the first and second set of values.
2. The method of claim 1 , further comprising labeling the first portion of the plurality of unknown values as the first set of values based on a supervised classification technique.
3. The method of claim 1 , further comprising labeling the first portion of the plurality of unknown values as the first set of values based on titles within the body of text.
4. The method of claim 1 , further comprising labeling the portion of the plurality of unknown attributes as the set of attributes based on a semi-supervised classification technique.
5. The method of claim 1 , further comprising labeling the portion of the plurality of unknown attributes as the set of attributes based on a unsupervised classification technique.
6. The method of claim 1 , further comprising:
determining, by an active learning component operatively coupled to the first classification sub-component and the second classification sub-component and implemented by the at least one processing device, at least one of a corrected attribute and a corrected value based on the set of attributes and the first and second set of values,
wherein updating the learning models further comprises updating the learning models based on either of the corrected attribute and the corrected value.
7. An apparatus, comprising at least one processing device, operable to identify at least one attribute and at least one value in a body of text comprising a plurality of unknown attributes and a plurality of unknown values, comprising:
a first classification sub-component, implemented by at least one processing device, operable to label, from a first portion of the body of text, a first portion of the plurality of unknown values as a first set of values; and
a second classification sub-component, implemented by the at least one processing device, operable to label, from a second portion of the body of text, a portion of the plurality of unknown attributes as a set of attributes and a second portion of the plurality of unknown values as a second set values,
wherein learning models implemented by the first classification sub-component and the second classification sub-component are updated, by the first classification sub-component and the second classification sub-component, based on the set of attributes and the first and second set of values.
8. The apparatus of claim 7 , wherein the first classification sub-component implements a supervised classification technique.
9. The apparatus of claim 7 , wherein the first classification sub-component is configured to operate on titles within the body of text.
10. The apparatus of claim 7 , wherein the second classification sub-component implements a semi-supervised classification technique.
11. The apparatus of claim 7 , wherein the second classification sub-component implements an unsupervised classification technique.
12. The apparatus of claim 7 , further comprising:
an active learning component, implemented by the at least one processing device and operatively connected to the first classification sub-component and the second classification sub-component, operable to provide at least one of a corrected attribute and a corrected value based on the set of attributes and the first and second set of values,
wherein the first classification sub-component and the second classification sub-component are further operable to update the learning models based on either of the corrected attribute and the corrected value.
13. A non-transitory computer-readable medium having stored thereon instructions that, when executed by a computer, cause the computer to:
in a body of text comprising a plurality of unknown attributes and a plurality of unknown values, label, based on a first classification technique operating on a first portion of the body of text, a first portion of the plurality of unknown values as a first set of values;
label, based on a second classification technique operating on a second portion of the body of text, a portion of the plurality of unknown attributes as a set of attributes and a second portion of the plurality of unknown values as a second set of values; and
update learning models used to implement the first classification technique and the second classification technique based on both the set of attributes and the first and second set of values.
14. The computer-readable medium of claim 13 , further comprising instructions that, when executed by the computer, cause the computer to label the first portion of the plurality of unknown values as the first set of values based on a supervised classification technique.
15. The computer-readable medium of claim 13 , further comprising instructions that, when executed by the computer, cause the computer to label the first portion of the plurality of unknown values as the first set of values based on titles within the body of text.
16. The computer-readable medium of claim 13 , further comprising instructions that, when executed by the computer, cause the computer to label the portion of the plurality of unknown attributes as the set of attributes based on a semi-supervised classification technique.
17. The computer-readable medium of claim 13 , further comprising instructions that, when executed by the computer, cause the computer to label the portion of the plurality of unknown attributes as the set of attributes based on a unsupervised classification technique.
18. The computer-readable medium of claim 13 , further comprising instructions that, when executed by the computer, cause the computer to:
determine at least one of a corrected attribute and a corrected value based on the set of attributes and the first and second set of values,
wherein the instructions are further operative to cause the computer to update the learning models based on either of the corrected attribute and the corrected value.