SYSTEMS AND METHODS FOR PROCESSING NEGATION IN NATURAL LANGUAGE SEARCH QUERIES
Systems and methods are described herein for processing negation in a natural language search query. The natural language search query is parsed to identify a plurality of portions of the natural language search query, and a first portion is determined to be a negation portion. A second portion to which the negation portion applies and a third portion to which the negation portion does not apply are then identified, and a content query is generated which includes a negation parameter corresponding to the second portion, and at least one other parameter corresponding to at least the third portion. Search results of the content query are then generated for output to the user.
1 . A method for processing negation in a natural language search query, the method comprising:
receiving the natural language search query;
parsing the natural language search query to generate a representation of the natural language search query;
identifying a negation portion within the representation of the natural language search query;
identifying one or more entity types within the representation of the natural language search query;
identifying a portion of the natural language search query to which the negation portion applies based at least in part on the one or more entity types identified within the representation of the natural language search query;
generating a content query that includes a negation parameter based on the identified portion of the natural language search query to which the negation portion applies;
generating search results based on the content query; and
outputting the search results.
2 . The method of claim 1 , wherein the identifying the portion of the natural language search query to which the negation portion applies comprises calculating distances between portions of the natural language search query and considering the identified one or more entity types.
3 . The method of claim 2 , wherein calculating distances between portions of the natural language search query and considering the identified one or more entity types comprises determining when different words refer to a same entity type within the natural language search query.
4 . The method of claim 1 , wherein the parsing the natural language search query to generate the representation of the natural language search query comprises parsing the natural language search query into one or more grammatical structures.
5 . The method of claim 4 , further comprising:
determining a context of the negation portion based at least in part on the one or more entity types identified within the representation of the natural language search query.
6 . The method of claim 5 , further comprising:
analyzing the one or more grammatical structures to determine grammatical relationships between words in the natural language search query, wherein:
the determining the context of the negation portion is based at least in part on the one or more entity types, or the identifying the portion of the natural language search query to which the negation portion applies is based at least in part on the determined grammatical relationships between the words in the natural language search query.
7 . The method of claim 1 , further comprising:
identifying parts of speech of the natural language search query with a conditional random field algorithm.
8 . The method of claim 1 , wherein the parsing the natural language search query comprises using a long short-term memory network.
9 . The method of claim 1 , further comprising:
determining a dialect of a language used in the natural language search query, wherein the determining the dialect of the language used in the natural language search query is performed before the identifying the negation portion within the representation of the natural language search query.
10 . The method of claim 1 , further comprising:
identifying a type of content corresponding to the natural language search query.
11 . A system for processing negation in a natural language search query, the system comprising:
a memory;
input/output circuitry configured to:
receive the natural language search query;
control circuitry configured to:
parse the natural language search query to generate a representation of the natural language search query;
identify a negation portion within the representation of the natural language search query;
identify one or more entity types within the representation of the natural language search query;
identify a portion of the natural language search query to which the negation portion applies based at least in part on the one or more entity types identified within the representation of the natural language search query;
generate a content query that includes a negation parameter based on the identified portion of the natural language search query to which the negation portion applies; and
generate search results based on the content query; and
wherein the input/output circuitry is further configured to:
output the search results.
12 . The system of claim 11 , wherein the control circuitry is configured to identify the portion of the natural language search query to which the negation portion applies by calculating distances between portions of the natural language search query and considering the identified one or more entity types.
13 . The system of claim 12 , wherein the control circuitry is configured to calculate distances between portions of the natural language search query and consider the identified one or more entity types by determining when different words refer to a same entity type within the natural language search query.
14 . The system of claim 11 , wherein the control circuitry is configured to parse the natural language search query to generate the representation of the natural language search query by parsing the natural language search query into one or more grammatical structures.
15 . The system of claim 14 , wherein the control circuitry is further configured to:
determine a context of the negation portion based at least in part on the one or more entity types identified within the representation of the natural language search query.
16 . The system of claim 15 , wherein the control circuitry is further configured to:
analyze the one or more grammatical structures to determine grammatical relationships between words in the natural language search query, wherein:
the determining the context of the negation portion is based at least in part on the one or more entity types, or the identifying the portion of the natural language search query to which the negation portion applies is based at least in part on the determined grammatical relationships between the words in the natural language search query.
17 . The system of claim 11 , wherein the control circuitry is further configured to:
identify parts of speech of the natural language search query with a conditional random field algorithm.
18 . The system of claim 11 , wherein the control circuitry is configured to parse the natural language search query by using a long short-term memory network.
19 . The system of claim 11 , wherein the control circuitry is further configured to:
determine a dialect of a language used in the natural language search query, wherein the determining the dialect of the language used in the natural language search query is performed before the identifying the negation portion within the representation of the natural language search query.
20 . The system of claim 11 , wherein the control circuitry is further configured to:
identify a type of content corresponding to the natural language search query.