RecLinkData.object {RecordLinkage}

Record Linkage Data Object

Description

S3 class representing information about record pairs for Record Linkage, as returned by functions compare.dedup and compare.linkage.

Value

A list with at least the following elements:

data (for type = "deduplication"):

Object of class "data.frame". Data frame of original records.

data1, data2 (for type = "linkage"):

Objects of class "data.frame". Data frames of original records.

pairs:

Object of class "data.frame" Data frame of data pairs. Each row represents the comparison pattern of two records, identified by columns id1 and id2. The other columns contain for each considered attribute a real number in the range [0..1] representing the degree of similarity. These columns are named according to the respective columns in data. The last column contains the matching status of the pair, coded as 1 for a match or 0 for a non-match.

frequencies:

Object of class "numeric" Numeric vector with average frequency of values for each column included in pairs (reciprocal of number of distinct values).

type:

Object of class "character" Identifies whether a linkage ("linkage") or a deduplication ("deduplication") project is represented.

.S3class:

Internal slot.

The following elements are optional:

M:

Object of class "numeric" Vector of m-probabilities as calculated by emWeights.

U:

Object of class "numeric" Vector of u-probabilities as calculated by emWeights.

W:

Object of class "numeric" Vector of log-likelihood weights as calculated by emWeights, corresponding to binary comparison patterns as created by bincombinations.

Wdata:

Object of class "numeric" Vector of log-likelihood weights as calculated by emWeights, corresponding to the rows of pairs.

Author(s)

Andreas Borg, Murat Sariyar

See Also

"RecLinkData" for the S4 representation. compare.dedup, which creates objects of this class. "RLBigData", an alternative data structure suitable for big data sets.