Full-text relevancy ranking
View Patent ↗A method and system for ranking relevancy of metadata associated with media on a computer network, such as multimedia and streaming media, include categorizing the metadata into sets of metadata. The categories are broad categories relating to areas such as who, what, when, and where, such as artist, media type, and creation date, creation location. Weights are assigned to each set of metadata. Weights are related to technical information such as bit rate, duration, sampling rate, frequency of occurrence of a specific term, etc. A score is calculated for ranking the relevancy of each set of metadata. The score is calculated in accordance with the assigned weight and category. This score is available for search systems (e.g., search engines) and/or users to determine the relative ranking of search results.
1. A computer-implemented method for ranking search results for multimedia items, the method comprising:
identifying metadata associated with at least one of a plurality of multimedia items, the metadata describing at least one semantic characteristic of the multimedia item;
comparing the identified metadata for the multimedia item to a reference database;
altering a field within the identified metadata based on the comparison;
determining an attribute for the multimedia item based on the identified metadata;
storing an indication that, for the multimedia item, the determined attribute has a value represented by the identified metadata;
receiving from a user a search query comprising at least one search term;
determining an attribute for the search term whose value is represented by the search term;
identifying at least one candidate multimedia item from the plurality of multimedia items for which an indication is stored that, for the candidate multimedia item, the determined attribute corresponds to the attribute determined for the search term determining a score for the at least one identified candidate multimedia item; and
providing instructions to display the identified candidate multimedia item in search results to the user based on the determined score.
2. The method of claim 1 further comprising ranking the identified candidate multimedia item according to the determined score.
3. The method of claim 1 further comprising assigning a weight to each search term based on an attribute determined for each search term.
4. The method of claim 3 , wherein the score is additionally based on the weights assigned to the one or more search terms in the search query.
5. The method of claim 1 , wherein the determined attribute is a who attribute.
6. The method of claim 5 , wherein the who attribute indicates an artist of the multimedia item.
7. The method of claim 1 , wherein the determined attribute is a what attribute.
8. The method of claim 7 , wherein the what attribute indicates a type of the multimedia item.
9. The method of claim 1 , wherein the determined attribute is a where attribute.
10. The method of claim 9 , wherein the where attribute indicates a creation location of the multimedia item.
11. The method of claim 1 , wherein the determined attribute is a when attribute.
12. The method of claim 11 , wherein the when attribute indicates a creation date of the multimedia item.
13. The method of claim 1 , further comprising:
identifying noisy metadata within the identified metadata;
comparing the noisy metadata to corresponding reference data from the reference database; and
replacing the noisy metadata with corresponding reference data from the reference database.
14. The method of claim 13 , wherein comparing the noisy metadata to corresponding reference data further comprises:
determining the number of common keywords in the noisy metadata and the corresponding reference data;
assigning a score to the corresponding reference data based on the number of common keywords;
adjusting the score based on weights associated with predetermined categories of keywords; and
selecting the corresponding reference data with the highest adjusted score.
15. The method of claim 1 , further comprising replacing connecting characters in the metadata with white space.
16. The method of claim 1 , wherein the reference database comprises multiple databases containing reference information for the plurality of multimedia items.
17. A computer-readable storage medium including program instructions that cause a computer system to rank search results for multimedia items by:
identifying metadata associated with at least one of a plurality of multimedia items, the metadata describing at least one semantic characteristic of the multimedia item;
comparing the identified metadata to a reference database;
replacing noisy metadata within the identified metadata with reference metadata from the reference database based on the comparison;
determining an attribute for the multimedia item based on the identified metadata;
storing an indication that, for the multimedia item, the determined attribute has a value represented by the identified metadata;
receiving from a user a search query comprising at least one search term;
determining an attribute for the search term whose value is represented by the search term;
identifying at least one candidate multimedia item from the plurality of multimedia items for which an indication is stored that, for the candidate multimedia item, the determined attribute corresponds to the attribute determined for the search term; and
determining a score for the at least one identified candidate multimedia item.
18. The computer-readable storage medium of claim 17 , wherein the metadata comprises elements related to content of the multimedia item.
19. The computer-readable storage medium of claim 17 , wherein the metadata comprises elements related to intellectual property rights associated with the multimedia item.
20. The computer-readable storage medium of claim 17 , wherein the metadata comprises elements related to instantiation of the multimedia item.
21. A method performed by a computer system to rank search results for multimedia items, the method comprising:
identifying metadata associated with at least one of a plurality of multimedia items, the metadata specifying a value that describes at least one semantic characteristic of the multimedia item;
comparing the identified metadata to a reference database;
altering a field within the identified metadata based on the comparison;
determining an attribute for the multimedia item based on the identified metadata;
storing an indication that, for the multimedia item, the determined attribute has a value represented by the identified metadata;
receiving from a user a search query comprising at least one search term;
determining an attribute for the search term whose value is represented by the search term;
identifying at least one candidate multimedia item from the plurality of multimedia items for which an indication is stored that, for the candidate multimedia item, the determined attribute corresponds to the attribute determined for the search term;
determining a score for the at least one identified candidate multimedia item; and
providing instructions to display the identified candidate multimedia item in search results to the user based on the determined score, wherein code implementing the method is stored in memory of the computer system for execution by a processor of the computer system.