Location context mining

An apparatus comprising:

• a computing device;
• and
• a memory configured to execute computer-readable instructions thereupon which, when executed by the computing device, cause the computing device to
• receive one or more documents;
• identify location words within the documents;
• identify candidate location context words within the documents, the candidate location context words being non-geographic words located within a specified word count of the location words with the documents, the candidate location context words being associated with the identified location words within the documents, where a set of all possible candidate location context words ci,j for each location word in a list of location words Xi,j being defined as:
• c
• i
• ,
• j
• =
• {
• w
• i
• ,
• j
• k
• ∈
• D
• i
• ❘
• /
• (
• X
• ij
• ,
• w
• ij
• k
• )
• ≤
• T
• }
• ,
• where wi,jk is the kth word in a document Di, the function /( . . . ) being a measure of a distance between the two words, and T being a predefined threshold;
• establish a set of possible pairings between the location words and the candidate location context words;
• and
• generate a score for each pairing comprising a term frequency inverse document frequency (TF-IDF) score.

A computer-implemented method for location context mining within document text, the computer-implemented method comprising performing computer-implemented operations for:

• receiving one or more documents, wherein the documents comprise travel content within the world wide web;
• identifying location words within the documents;
• identifying candidate location context words within the documents, the candidate location context words being non-geographic words located within a specified word count of the location words with the documents, the candidate location context words being associated with the identified location words within the documents, where a set of all possible candidate location context words for each location word in a list of location words Xi,j being defined as:
• c
• i
• ,
• j
• =
• {
• w
• i
• ,
• j
• k
• ∈
• D
• i
• ❘
• /
• (
• X
• ij
• ,
• w
• ij
• k
• )
• ≤
• T
• }
• ,
• where wi,jk is the kth word in a document Di, the function /( . . . ) being a measure of a distance between the two words, and T being a predefined threshold;
• establishing a set of possible pairings between the location words and the candidate location context words;
• generating a relevance score for each pairing;
• supporting a query of the location words based on one of the candidate location context words;
• and
• supporting a query of the candidate location context words based on one of the location words.