Organizing speech search results
A method, system, and phonetic search engine are described that enable phonetic searches to have increased relevancy to the searcher. Specifically, phonetic searches on a database containing phonetically-searchable content can have one or more phonetically-confusable terms included therein, thereby creating search results that more faithfully reflect the search terms used during the phonetic search of the database.
1. A phonetic search engine, comprising:
a processor; and a computer-readable storage medium storing computer readable instructions which, when executed by the processor, cause the processor to perform steps comprising:
receiving one or more search terms for a phonetic search to be performed on a database containing phonetically-searchable content;
receiving one or more undesired terms associated with the phonetic search;
generating a database query including at least one of the received one or more search terms and one or more undesired terms;
transmitting the database query to the database;
receiving a set of search results from the database in response to the database query;
deriving one or more relative confidence metrics from a determined phonetic distance metrics which are the phonetic distances from each search result to one or more search terms, and the phonetic distance from each search result to one or more undesired terms;
organizing the set of search results for presentation based on the derived one or more relative confidence metrics;
one or more search results from the set of search results are not displayed when the one or more search results are phonetically closer to an undesired term of the undesired terms than any one of the one or more search terms; and
displaying the organized set of search results on a user interface.
2. The phonetic search engine of claim 1 , wherein a first search result in the set of search results comprises a first relative confidence metric associated therewith, wherein a second search result in the set of search results comprises a second relative confidence metric associated therewith, wherein the first confidence metric is greater than the second confidence metric, and wherein the steps further comprise causing the first search result to be presented via a search engine interface in a more prominent position than the second search result.
3. The phonetic search engine of claim 2 , wherein the first relative confidence metric is greater than the second confidence metric due to the first search result having a shorter phonetic distance than the second search result to a desired search term in the one or more search terms.
4. The phonetic search engine of claim 2 , wherein the one or more search terms comprises a desired search term and an undesired search term, wherein the undesired search term is phonetically similar to the desired search term, wherein the first search result is a first phonetic distance away from the desired search term and a second phonetic distance away from the undesired search term, wherein the second search result is a third phonetic distance away from the desired search term and a fourth phonetic distance away from the undesired search term.
5. The phonetic search engine of claim 4 , wherein the first relative confidence metric is determined by calculating an inverse ratio of the first phonetic distance and second phonetic distance and wherein the second relative confidence metric is determined by calculating an inverse ratio of the third phonetic distance and the fourth phonetic distance.
6. The phonetic search engine of claim 4 , wherein the database query comprises a “BUT NOT” operator between the desired search term and the undesired search term.
7. The phonetic search engine of claim 1 , wherein the one or more relative confidence metrics for each result in the set of search results accounts for relative phonetic distances between search results in the set of search results.
8. The phonetic search engine of claim 1 , wherein the phonetic distance metrics used to derive the one or more relative confidence metrics include one or more of a number of common syllables, a sequence of common syllables, a number of common phonemes, a sequence of common phonemes, a number of different syllables, and a number of different phonemes.
9. The phonetic search engine of claim 1 , wherein the phonetically-searchable content comprises at least one of audio and video content from a communication session between two or more session participants.
10. The phonetic search engine of claim 1 , wherein the presentation of the set of search results includes a clustering of search results that are phonetically clustered with one another.
11. A system, comprising:
one or more servers configured to provide an interface between a user communication device and a database, wherein the database contains phonetically-searchable content, wherein the one or more servers include a search engine, wherein the search engine is configured to
receive one more search terms for a phonetic search to be performed on the database, wherein the one or more servers comprise a processor, wherein the processor;
retrieves a set of search results from the database in response to a database query, wherein the database query comprises one or more search terms and one or more undesired terms for a phonetic search to be performed on the database;
derives one or more relative confidence metrics from a determined phonetic distance metrics which are the phonetic distances from each search result to one or more search terms, and the phonetic distance from each search result to one or more undesired terms;
organizes the set of search results retrieved from the database based on the derived one or more relative confidence metrics,
one or more search results from the set of search results are not displayed when the one or more search results are phonetically closer to an undesired term of the undesired terms than any one of the one or more search terms; and
displaying the organized set of search results on a user interface.
12. The system of claim 11 , wherein the one or more relative confidence metrics represent a confidence that a search result is closer to a desired search term than an undesired search term.
13. The system of claim 12 , wherein the database query comprises a “BUT NOT” operator associating the desired search term and the undesired search term.
14. The system of claim 11 , wherein the phonetic distance metrics include a phonetic distance between the search result and the desired search term and a phonetic distance between the search result and the undesired search term.
15. The system of claim 14 , wherein the phonetic distance metrics further include a phonetic distance between the search result and at least one other search result in the set of search results.
16. The system of claim 14 , wherein a relative confidence metric for the search result comprises an inverse ratio of the phonetic distance between the search result and the desired term and the phonetic distance between the search result and the undesired search term.
17. The system of claim 14 , wherein phonetic distances are determined, at least in part, by considering one or more of a number of common syllables, a sequence of common syllables, a number of common phonemes, a sequence of common phonemes, a number of different syllables, and a number of different phonemes.
18. The system of claim 11 , wherein the phonetically-searchable content comprises at least one of audio and video content from a communication session between two or more session participants.
19. A method of performing phonetic searches, the method comprising:
receiving, by a processor, one or more search terms for a phonetic search to be performed on a database containing phonetically-searchable content;
receiving, by the processor, one or more undesired terms associated with the phonetic search;
generating, by the processor, a database query that includes at least one of the received one or more search terms and one or more undesired terms;
transmitting, by the processor, the database query to a database interface;
receiving, by the processor, a set of search results from the database interface in response to the database query;
deriving, by the processor, one or more relative confidence metrics from a determined phonetic distance metrics which are the phonetic distances from each search result to one or more search terms, and the phonetic distance from each search result to one or more undesired terms;
organizing, by the processor, the set of search results based on the derived one or more relative confidence metrics;
one or more search results from the set of search results are not displayed when the one or more search results are phonetically closer to an undesired term of the undesired terms than any one of the one or more search terms; and
displaying the organized set of search results on a user interface.
20. The method of claim 19 , further comprising:
calculating, by the processor, a phonetic distance to a desired search term included in the database query for each result in the set of search results;
calculating, by the processor, a phonetic distance to an undesired search term included in the database query for each result in the set of search results;
determining, by the processor, an inverse ratio of the phonetic distance to the desired search term and the phonetic distance to the undesired search term for each result in the set of search results; and
including, by the processor, the inverse ratio in the one or more relative confidence metrics.