Query classification based on search engine results

摘要

The technology described relates to performing query classification by using search engine results. In one aspect, a method includes identifying, from search results responsive to first search queries, first collections of features associated with classifications for the first search queries; in response to receiving a second search query, generating from search results responsive to the second search query a second collection of features associated with the second search query; and comparing the first collections of features to the second collection of features to determine a particular classification for the second search query.

主权项

1. A computer-implemented method comprising:
defining a plurality of classifications; selecting a plurality of training queries as representative of a particular classification of the plurality of classifications; obtaining search results responsive to respective training queries in the plurality of training queries selected as representative of the particular classification; generating a classification feature vector for the particular classification from the search results responsive to respective training queries in the plurality of training queries selected as representative of the particular classification, wherein corresponding elements of the classification feature vector represent corresponding terms, and wherein a value of an element of the classification feature vector for the particular classification represents a term frequency of a corresponding term in the obtained search results responsive to the training queries selected for the particular classification; receiving a search query; obtaining search results responsive to the search query; generating a search result vector from the search results responsive to the search query; determining that the search result vector is closest to the classification feature vector for the particular classification relative to other respective classification feature vectors generated for other classifications in the plurality of classifications; and in response to determining that the search result vector is closest to the classification feature vector for the particular classification relative to other respective classification feature vectors generated for other classifications in the plurality of classifications, obtaining search results having a vertical content type according to the particular classification from a particular corpus; and providing the obtained search results having the vertical content type in response to the search query.