Method and apparatus for determining search result demographics
Techniques are presented for producing demographics, in an automated fashion, from a search result of computer-accessible content. While the demographics can be determined for a research product that has been produced by any technique, they are particularly useful when applied to an automated frame-based search approach. Frame-based search engines are presented for technology profiling, healthcare-related search and brand research. Determination of a demographic proceeds at two levels: member and population. At the member level, a demographic characteristic can be determined applicable with either total or partial certainty. Each value assigned by a demographic, to a population member, has a confidence level associated with it and the assignments can be represented by a Confidence Distribution. Summarization of a demographic, at the population level, depends upon whether the certainty assignments, at the member level, are total or partial. Declarant Demographics are presented. Approaches, to determining Declarant Demographics, are presented.
1. A method for frame-based search, performed by computing hardware and programmable memory, comprising the following steps:
receiving a first rule, for producing an instance in accordance with a first frame, wherein the first frame comprises a first input role and a first output role;
receiving a second rule, for producing an instance in accordance with a second frame, wherein the second frame comprises a second input role and a second output role;
receiving a first source corpus;
identifying first and second units of natural language from, respectively, first and second snippets of the first source corpus;
producing a first instance, from the first unit, by application of the first rule, wherein the first instance has a first input value assigned to its first input role and a first output value assigned to its first output role;
producing a second instance, from the second unit, by application of the second rule, wherein the second instance has a second input value assigned to its second input role and a second output value assigned to its second output role;
matching an input user query to the first input value of the first instance, and the second input value of the second instance;
determining the first and second instances represent a same third frame;
determining a same third value as representative of the first and second input values;
determining a same fourth value as representative of the first and second output values;
producing, as a result of computing hardware and programmable memory, a first result that contains a first result-value and a first result-base, wherein the first result-value is the fourth value and the first result-base contains the first and second snippets;
determining, as a result of computing hardware and programmable memory, a member-level demographic value for each member of the first result-base;
determining, as a result of computing hardware and programmable memory, a first demographic value, for the first result, by combining the member-level demographic values; and
displaying, to a user as part of a search result, the first demographic value as a demographic determined for the first result-value.
2. The method of claim 1 , wherein the step of determining a member-level demographic value for each member of the first result-base further comprises:
determining whether a first lexical unit, with a demographic association, is present.
3. The method of claim 2 , wherein the first lexical unit is indicative of a geographical area.
4. The method of claim 2 , wherein the lexical unit is produced by application of automated machine learning procedures to training corpora.
5. The method of claim 1 , wherein the step of determining a member-level demographic value for each member of the first result-base further comprises:
application of a third linguistic rule, that triggers upon a logical form, and, if triggered, has an action that indicates a presence of a demographic characteristic.
6. The method of claim 5 , wherein the third linguistic rule tests for self-referential demographic identification.
7. The method of claim 2 , wherein the third linguistic rule requires a logical form with a verb, actor, and undergoer.
8. The method of claim 2 , wherein the third linguistic rule requires a logical form with a verb of a form of to be, and a self-referential actor.
9. The method of claim 2 , wherein the third linguistic rule, if triggered, indicates the presence of the demographic characteristic with a confidence distribution.
10. The method of claim 1 , further comprising:
producing a first Logical Form semantic representation for the first unit of natural language of the first corpus;
determining whether a first conditional part of the first rule matches the first Logical Form;
producing, if the first conditional part matches, in accordance with a first action part of the first rule, an assignment of values of the first Logical Form to the first input role and first output role of the first instance;
producing a second Logical Form semantic representation for the second unit of natural language of the first corpus;
determining whether a second conditional part of the second rule matches the second Logical Form; and
producing, if the second conditional part matches, in accordance with a second action part of the second rule, an assignment of values of the second Logical Form to the second input role and second output role of the second instance.