Searching for information based on generic attributes of the query

• US 8,560,513 B2
• Filed: 09/07/2010
• Issued: 10/15/2013
• Est. Priority Date: 09/27/2009
• Status: Expired due to Fees

• Current Assignee: Alibaba Group Holding Ltd.
• Original Assignee: Alibaba Group Holding Ltd.
• Inventors: Lin, Feng, Kang, Zhanhui, Huang, Peng
• Primary Examiner(s): Ortiz Ditren, Belix M
• Application Number:
  US12/807,505
• Publication Number:
  US 20110078127A1
• Time in Patent Office: 1,134 Days
• Field of Search: 707/706
• US Class Current: 707/706
• CPC Class Codes: G06F 16/23 Updating
• CPC Class Codes: G06F 16/24575 using context
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 16/3322 using system suggestions G06F16/3325 takes precedence
• CPC Class Codes: G06F 16/3323 using document space presentation or visualization, e.g. category, hierarchy or range presentation and selection
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 16/9538 Presentation of query results
• CPC Class Codes: H04L 67/01 Protocols

Abstract

Searching information includes: receiving current query data from a client; extracting generic attribute features of the current query data, wherein the generic attribute features are used for calculating a plurality of confidence degrees of the current query data that correspond to a plurality of categories, each of the confidence degrees indicating a degree of confidence that the current query data belongs to a respective one of the plurality of categories; determining the plurality of confidence degrees of the current query data based at least in part on the generic attribute features; searching in a chosen category for a search result that corresponds to the current query data, the chosen category being one of the plurality of categories and being chosen based at least in part on the plurality of confidence degrees; and returning the search result.

Claims

1. A method for searching information, comprising:

receiving current query data from a client;

extracting generic attribute features of the current query data, wherein the generic attribute features are used for calculating a plurality of confidence degrees of the current query data that correspond to a plurality of categories, each of the confidence degrees indicating a degree of confidence that the current query data belongs to a respective one of the plurality of categories;

determining the plurality of confidence degrees of the current query data based at least in part on the generic attribute features, wherein the determining of the plurality of confidence degrees is based on a maximum entropy model, a linear regression, or a support vector machine model;

selecting a category based at least in part on the plurality of confidence degrees, the selected category being one of the plurality of categories and having a confidence degree higher than a confidence degree of another category;

searching in the selected category for a search result that corresponds to the current query data; and

returning the search result.

2. The method of claim 1, wherein the generic attribute features comprise a word-level feature that corresponds to a keyword in the current query data.

3. The method of claim 1, wherein the generic attribute features comprise a syntax feature that corresponds to an etymon corresponding to a headword in the current query data.

4. The method of claim 3, wherein extracting the generic attribute features including the syntax feature comprises:

performing part-of-speech tagging of a keyword in the current query data and identifying a phrase including the keyword in the current query data;

obtaining a noun phrase in the current query data based at least in part on results of the part-of-speech tagging and the identification of the phrase;

extracting the etymon associated with the headword of the noun phrase.

5. The method of claim 1, wherein the generic attribute features comprise a semantic feature that corresponds to a semantic tag corresponding to the keyword.

6. The method of claim 5, wherein extracting the generic attribute features including the semantic feature comprises:

identifying the type of the keyword of the current query data; and

acquiring an action verb tag, a product tag, a brand tag, a model tab, or another tag of the keyword according to the identified type.

7. The method of claim 1, wherein the generic attribute features comprise a click feature that corresponds to a clicked category corresponding to the current query data in a click log.

8. The method of claim 7, wherein extracting the generic attribute features including the click feature comprises:

determining whether the current query data is present in the click log, wherein the click log is adapted to store a correspondence relationship between the query data and the click feature;

in the event that the current query data is present in the click log, acquiring a first predetermined number of clicked categories corresponding to the current query data from the preset click log, wherein the clicked categories are categories to which a corresponding product clicked by a user after searching for the query data belongs; and

in the event that the current query data is not present in the click log, inquiring in the click log about clicked query data closest to the current query data, wherein the clicked query data is at the shortest edit distance in the click log from the current query data, and acquiring the first predetermined number of categories corresponding to the clicked query data as the click feature.

9. The method of claim 1, wherein the generic attribute features comprise a selection feature that corresponds to at least a selected category corresponding to the current query data in a selection log.

10. The method of claim 9, wherein extracting the generic attribute features including the selection feature comprises:

determining whether the current query data is present in the selection log, wherein the selection log is adapted to store a correspondence relationship between the query data and the selection feature;

in the event that the current query data is present in the selection log, acquiring a first predetermined number of selected categories corresponding to the current query data from the selection log, wherein the selected categories represent categories associated with the current query data in a mapping table of a search engine; and

in the event that the current query data is not present in the selection log, inquiring in the selection log about selected query data closest to the current query data, wherein, the selected query data is at the shortest edit distance in the selection log from the current query data, and acquiring the first predetermined number of categories corresponding to the selected query data as selection features of the current query data.

11. The method of claim 1, wherein the generic attribute features comprise a session feature that corresponds to a substantially optimal category corresponding to a previous one query preceding the current query data in the same session.

12. The method of claim 11, wherein extracting the generic attribute features including the session feature comprises:

acquiring preceding query data that precedes the current query data in the same session;

acquiring a category to which an optimal categorizing result corresponding to the preceding query data belongs; and

setting the category to which the optimal categorizing result belongs as a session feature of the current query data.

13. The method of claim 1, wherein determining the plurality of confidence degrees of the current query data based at least in part on the generic attribute features includes:

setting the generic attribute features as given information of a conditional probability model;

calculating a plurality of posterior probabilities that the current query belongs to the respective categories under condition of the given information using the conditional probability model; and

setting the plurality of posterior probabilities as the plurality of confidence degrees.

14. The method of claim 1, further comprising:

identifying selected ones of the plurality of confidence degrees that meet a preset first threshold and their corresponding categories; and

returning the selected ones of the plurality of confidence degrees and their corresponding categories to the client for selecting the selected category for the current query data.

15. The method of claim 1, wherein the selected category has a maximum calculated confidence degree for the current query data.

18. The method of claim 1, wherein the selecting of the category includes selecting a predetermined number of categories of the plurality of categories.

19. The method of claim 1, wherein the selecting of the category includes selecting one of the plurality of categories exceeding a predefined threshold.

16. A system for searching information, comprising:

one or more processors configured to;

receive current query data from a client;

extract generic attribute features of the current query data, wherein the generic attribute features are used for calculating a plurality of confidence degrees of the current query data that correspond to a plurality of categories, each of the confidence degrees indicating a degree of confidence that the current query data belongs to a respective one of the plurality of categories;

determine the plurality of confidence degrees of the current query data based at least in part on the generic attribute features, wherein the determining of the plurality of confidence degrees is based on a maximum entropy model, a linear regression, or a support vector machine model;

select a category based at least in part on the plurality of confidence degrees, the selected category being one of the plurality of categories and having a confidence degree higher than a confidence degree of another category;

search in the selected category for a search result that corresponds to the current query data; and

return the search result; and

one or more memories coupled to the one or more processors, configured to provide the processors with instructions.

17. A computer program product for searching information, the computer program product being embodied in a tangible non-transitory computer readable storage medium and comprising computer instructions for:

receiving current query data from a client;

extracting generic attribute features of the current query data, wherein the generic attribute features are used for calculating a plurality of confidence degrees of the current query data that correspond to a plurality of categories, each of the confidence degrees indicating a degree of confidence that the current query data belongs to a respective one of the plurality of categories;

determining the plurality of confidence degrees of the current query data based at least in part on the generic attribute features, wherein the determining of the plurality of confidence degrees is based on a maximum entropy model, a linear regression, or a support vector machine model;

selecting a category based at least in part on the plurality of confidence degrees, the selected category being one of the plurality of categories and having a confidence degree higher than a confidence degree of another category;

searching in the selected category for a search result that corresponds to the current query data; and

returning the search result.