triples: Bind Triples to a Textmatrix

In lsa: Latent Semantic Analysis

Bind Triples to a Textmatrix

Description

Allows to store, manage and retrieve SPO-triples (subject, predicate, object) bound to the document columns of a document term matrix.

Usage

getTriple( M, subject, predicate )
setTriple( M, subject, predicate, object )
delTriple( M, subject, predicate, object )
getSubjectId( M, subject )

Arguments

M	the document term matrix (see textmatrix).
subject	column number or column name (e.g., "doc3" or 3).
predicate	predicate of the triple sentence (e.g., "has\_category" or "has\_grade").
object	value of the triple sentence (e.g., "14" or 14).

Details

SPO-Triples are simple facts of the uniform structure (subject, predicate, object). A subject is typically a document in the given document-term matrix M, i.e. its document title (as in the column names) or its column position. A key-value pair (the predicate and the object) can be bound to this subject.

This can be used, for example, to store classification information about the documents of the text base used.

The triple data is stored in the environment of M constructed by textmatrix().

Whenever a matrix has to be used which has not been generated by this function, its class should be set to 'textmatrix' and an environment has to be added manually via:

class(mymatrix) = "textmatrix"

environment(mymatrix) = new.env()

Alternatively, as.matrix() can be used to convert a matrix to a textmatrix. To spare memory, the manual method might be of advantage.

In getTriple(), the arguments subject and predicate are optional.

Value

textmatrix

the document-term matrix (including row and column names).

Author(s)

Fridolin Wild [e-mail]

See Also

textmatrix

Examples

    x = matrix(2,2,3) # we fake a document term matrix
    rownames(x) = c("dog","mouse") # fake term names
    colnames(x) = c("doc1","doc2","doc3") # fake doc titles
    class(x) = "textmatrix" # usually done by textmatrix()
    environment(x) = new.env() # usually done by textmatrix()
    
    setTriple(x, "doc1", "has_category", "15")
    setTriple(x, "doc2", "has_category", "7")
    setTriple(x, "doc1", "has_grade", "5")
    setTriple(x, "doc1", "has_category", "11")
    
    getTriple(x, "doc1")
    getTriple(x, "doc1")[[2]]
    getTriple(x, "doc1", "has_category") # -> [1] "15" "11"
    
    delTriple(x, "doc1", "has_category", "15")
    getTriple(x, "doc1", "has_category") # -> [1] "11"
    

lsa documentation built on May 9, 2022, 9:10 a.m.