Package {automatedRecLin}

Title:	Record Linkage Based on an Entropy-Maximizing Classifier
Version:	1.1.0
Description:	The goal of 'automatedRecLin' is to perform record linkage (also known as entity resolution) in unsupervised or supervised settings. It compares pairs of records from two datasets using selected comparison functions to estimate the probability or density ratio between matched and non-matched records. Based on these estimates, it predicts a set of matches that maximizes entropy. For details see: Lee et al. (2022) https://www150.statcan.gc.ca/n1/pub/12-001-x/2022001/article/00007-eng.htm, Vo et al. (2023) https://ideas.repec.org/a/eee/csdana/v179y2023ics0167947322002365.html, Sugiyama et al. (2008) <doi:10.1007/s10463-008-0197-x>.
License:	GPL-3
Encoding:	UTF-8
URL:	https://github.com/ncn-foreigners/automatedRecLin, http://ncn-foreigners.ue.poznan.pl/automatedRecLin/
BugReports:	https://github.com/ncn-foreigners/automatedRecLin/issues
Imports:	blocking, data.table, densityratio, FixedPoint, methods, nleqslv, purrr, reclin2, stats, utils
Suggests:	knitr, rmarkdown, tinytest, xgboost
Depends:	R (≥ 4.1.0)
LazyData:	true
VignetteBuilder:	knitr
Config/roxygen2/version:	8.0.0
NeedsCompilation:	no
Packaged:	2026-05-07 16:41:23 UTC; adamstruzik
Author:	Adam Struzik [aut, cre], Maciej Beręsewicz [aut, ctb]
Maintainer:	Adam Struzik <adastr5@st.amu.edu.pl>
Repository:	CRAN
Date/Publication:	2026-05-08 07:40:21 UTC

A_example dataset

Description

An example dataset containing artificial personal data.

Usage

A_example

Format

A data.frame with 10 records. Each row represents one record, with the following columns: name, surname, and city. Some records can be matched with records in the B_example dataset.

Examples

data("A_example")
A_example

B_example dataset

Description

An example dataset containing artificial personal data.

Usage

B_example

Format

A data.frame with 12 records. Each row represents one record, with the following columns: name, surname, and city. Some records can be matched with records in the A_example dataset.

Examples

data("B_example")
B_example

Absolute Distance Comparison Function

Description

Creates a function that calculates the absolute distance between two values.

Usage

abs_distance()

Value

Returns a function taking two arguments, x and y, and returning their absolute difference.

Author(s)

Adam Struzik

Examples

cmp <- abs_distance()
cmp(1, 5) # returns 4

Fictional census data

Description

This dataset was created by Paula McLeod, Dick Heasman and Ian Forbes, ONS, for the ESSnet DI on-the-job training course, Southampton, 25-28 January 2011. It contains fictional data representing some observations from a decennial Census.

Usage

census

Format

A data.table with 25343 records. Each row represents one record, with the following columns:

• person_id – a unique number for each person, consisting of postcode, house number and person number,

• pername1 – forename,

• pername2 – surname,

• sex – gender (M/F),

• dob_day – day of birth,

• dob_mon – month of birth,

• dob_year – year of birth,

• hse_num – house number, a numeric label for each house within a street,

• enumcap – an address consisting of house number and street name,

• enumpc – postcode,

• str_nam – street name of person's household's street,

• cap_add – full address, consisting of house number, street name and postcode,

• census_id – person ID with "CENS" added in front.

References

McLeod, P., Heasman, D., Forbes, I. (2011). Simulated data for the ESSnet DI on-the-job training course, Southampton, 25-28 January 2011. https://wayback.archive-it.org/12090/20231221144450/https://cros-legacy.ec.europa.eu/content/job-training_en

Examples

data("census")
head(census)

Fictional customer data

Description

This dataset was created by Paula McLeod, Dick Heasman and Ian Forbes, ONS, for the ESSnet DI on-the-job training course, Southampton, 25-28 January 2011. It contains fictional observations from Customer Information System, which is combined administrative data from the tax and benefit systems.

Usage

cis

Format

A data.table with 24613 records. Each row represents one record, with the following columns:

• person_id – a unique number for each person, consisting of postcode, house number and person number,

• pername1 – forename,

• pername2 – surname,

• sex – gender (M/F),

• dob_day – day of birth,

• dob_mon – month of birth,

• dob_year – year of birth,

• enumcap – an address consisting of house number and street name,

• enumpc – postcode,

• cis_id – person ID with "CIS" added in front.

References

McLeod, P., Heasman, D., Forbes, I. (2011). Simulated data for the ESSnet DI on-the-job training course, Southampton, 25-28 January 2011. https://wayback.archive-it.org/12090/20231221144450/https://cros-legacy.ec.europa.eu/content/job-training_en

Examples

data("cis")
head(cis)

Create Comparison Vectors for Record Linkage

Description

Creates comparison vectors between records in two datasets based on specified variables and comparison functions.

Usage

comparison_vectors(
  A,
  B,
  variables,
  comparators = NULL,
  pairs = NULL,
  matches = NULL
)

Arguments

Details

Consider two datasets: A and B. For each pair of records (a,b) \in \Omega, the function creates a comparison vector \pmb{\gamma}_{ab} = (\gamma_{ab}^1,\gamma_{ab}^2,\ldots,\gamma_{ab}^K)' based on specified K variables and comparison functions.

Value

Returns a list containing:

• Omega – a data.table with comparison vectors between records from both datasets, including optional match information,

• variables – a character vector of key variables used for comparison,

• comparators – a list of functions used to compare pairs of records,

• match_prop – proportion of matches in the smaller dataset.

Note

Each comparison function must return another function, which serves as the actual comparator.

Author(s)

Adam Struzik

Examples

df_1 <- data.frame(
"name" = c("John", "Emily", "Mark", "Anna", "David"),
"surname" = c("Smith", "Johnson", "Taylor", "Williams", "Brown")
)
df_2 <- data.frame(
  "name" = c("Jon", "Emely", "Marc", "Michael"),
  "surname" = c("Smitth", "Jonson", "Tailor", "Henderson")
)
comparators <- list("name" = jarowinkler_complement(),
                    "surname" = jarowinkler_complement())
matches <- data.frame("a" = 1:3, "b" = 1:3)
result <- comparison_vectors(A = df_1, B = df_2, variables = c("name", "surname"),
                             comparators = comparators, matches = matches)
result

Controls for the kliep() Function

Description

Controls for the kliep() function used in the package.

Usage

control_kliep(scale = NULL, progressbar = FALSE, nfold = 2, ...)

Arguments

Value

Returns a list with parameters.

Author(s)

Adam Struzik

Create a Custom Record Linkage Model

Description

Creates a supervised record linkage model using a custom machine learning (ML) classifier.

Usage

custom_rec_lin_model(ml_model, vectors)

Arguments

Details

The custom_rec_lin_model() function creates a custom record linkage model, based on known matches and non-matches (which might later serve as a classifier for pairs outside training data). The procedure of creating a custom model based on training data is as follows.

1. Use comparison_vectors() to compare pairs of records.

2. Train a machine learning classifier using the Omega element of the output of comparison_vectors(). The classifier should predict the probability of matching based on a given vector.

3. Use custom_rec_lin_model() with appropriate arguments.

Value

Returns a list containing:

• b_vars – here NULL,

• cpar_vars – here NULL,

• cnonpar_vars – here NULL,

• b_params – here NULL,

• cpar_params – here NULL,

• cnonpar_params – here NULL,

• ratio_kliep – here NULL,

• ratio_kliep_list – here NULL,

• ml_model – ML model used for creating the record linkage model,

• pi_est – a prior probability of matching,

• match_prop – proportion of matches in the smaller dataset,

• variables – a character vector of key variables used for comparison,

• comparators – a list of functions used to compare pairs of records,

• methods – here NULL,

• prob_ratio – here "2".

Author(s)

Adam Struzik

Examples

if (requireNamespace("xgboost", quietly = TRUE)) {
  df_1 <- data.frame(
    "name" = c("James", "Emma", "William", "Olivia", "Thomas",
    "Sophie", "Harry", "Amelia", "George", "Isabella"),
    "surname" = c("Smith", "Johnson", "Brown", "Taylor", "Wilson",
    "Davis", "Clark", "Harris", "Lewis", "Walker")
  )
  df_2 <- data.frame(
    "name" = c("James", "Ema", "Wimliam", "Olivia", "Charlotte",
    "Henry", "Lucy", "Edward", "Alice", "Jack"),
    "surname" = c("Smith", "Johnson", "Bron", "Tailor", "Moore",
    "Evans", "Hall", "Wright", "Green", "King")
  )
  comparators <- list("name" = jarowinkler_complement(),
                      "surname" = jarowinkler_complement())
  matches <- data.frame("a" = 1:4, "b" = 1:4)
  vectors <- comparison_vectors(A = df_1, B = df_2, variables = c("name", "surname"),
                               comparators = comparators, matches = matches)
  model_xgb <- xgboost::xgboost(x = as.matrix(vectors$Omega[, c("gamma_name", "gamma_surname")]),
                       y = factor(vectors$Omega$match),
                       objective = "binary:logistic", eval_metric = "logloss",
                       nrounds = 100, verbosity = 0, nthread = 1)
  custom_xgb_model <- custom_rec_lin_model(model_xgb, vectors)
  custom_xgb_model
}

Parametric Bootstrap for Standard Error Estimation in MEC

Description

WORK IN PROGRESS

This function is currently under development.

Usage

est_se_bootstrap(mec_result, B = 100, alpha = 0.05)

Arguments

Author(s)

Adam Struzik

Jaro-Winkler Distance

Description

Creates a function that calculates the Jaro-Winkler distance between two strings, defined as 1 - \text{Jaro-Winkler similarity}.

Usage

jarowinkler_complement()

Value

Returns a function taking two string arguments, x and y, and returning the Jaro-Winkler distance.

Author(s)

Adam Struzik

Unsupervised Maximum Entropy Classifier for Record Linkage

Description

Implements several extensions to the maximum entropy classification (MEC) algorithm for record linkage (see Lee et al. (2022)), iteratively estimating probability/density ratios to classify record pairs into matches and non-matches based on comparison vectors.

Usage

mec(
  A,
  B,
  variables,
  comparators = NULL,
  methods = NULL,
  duplicates_in_A = FALSE,
  start_params = NULL,
  nonpar_hurdle = TRUE,
  set_construction = NULL,
  target_rate = 0.03,
  max_iter_bisection = 100,
  tol = 0.005,
  delta = 0.5,
  eps = 0.05,
  max_iter_em = 10,
  tol_em = 1,
  controls_nleqslv = list(),
  controls_kliep = control_kliep(),
  true_matches = NULL,
  verbose = FALSE
)

Arguments

Details

Consider two datasets without duplicates: A and B. Let the bipartite comparison space \Omega = A \times B consist of matches M and non-matches U between the records in files A and B. For any pair of records (a,b) \in \Omega, let \pmb{\gamma}_{ab} = (\gamma_{ab}^1,\gamma_{ab}^2, \ldots,\gamma_{ab}^K)' be the comparison vector between a set of key variables. The original MEC algorithm uses the binary comparison function to evaluate record pairs across two datasets. However, this approach may be insufficient when handling datasets with frequent errors across multiple variables.

We propose the use of continuous comparison functions to address the limitations of binary comparison methods. We consider every semi-metric, i.e., a function d: A \times B \to \mathbb{R}, satisfying the following conditions:

1. d(x,y) \geq 0,

2. d(x,y) = 0 if and only if x = y,

3. d(x,y) = d(y,x).

For example, we can use the Jaro-Winkler distance for character variables (which is implemented in the automatedRecLin package as jarowinkler_complement()) or the Euclidean distance for numerical variables. The automatedRecLin package allows the use of a different comparison function for each key variable (which should be specified as a list in the comparators argument). The default function for each key variable is cmp_identical() (the binary comparison function).

The mec() function offers different approaches to estimate the probability/density ratio between matches and non-matches, which should be specified as a list in the methods argument. The available methods suitable for the binary comparison function are "binary" and "hit_miss". Both assume that \gamma_{ab}^k|M and \gamma_{ab}^k|U follow Bernoulli distributions. "binary" and "hit_miss" both estimate the parameters for the matches iteratively, but "binary" estimates the parameters for the non-matches only at the start, while "hit_miss" does so iteratively using a hit-miss model (for details see Lee et al. (2022)). "binary" is the default method for each variable.

For the continuous semi-metrics we suggest the usage of "continuous_parametric" or "continuous_nonparametric" method. The "continuous_parametric" method assumes that \gamma_{ab}^k|M and \gamma_{ab}^k|U follow hurdle Gamma distributions. The density function of a hurdle Gamma distribution is characterized by three parameters p_0 \in (0,1) and \alpha, \beta > 0 as follows:

f(x;p_0,\alpha,\beta) = p_0^{\mathbb{I}(x = 0)}[(1 - p_0) v(x;\alpha,\beta)]^{\mathbb{I}(x > 0)},

where

v(x;\alpha,\beta) = \frac{\beta^{\alpha} x^{\alpha - 1} \exp(-\beta x)} {\Gamma(\alpha)}

is the density function of a Gamma distribution (for details see Vo et al. (2023)). At the beginning, the algorithm estimates the parameters for the non-matches and then does it iteratively for the matches. The "continuous_nonparametric" method does not assume anything about the distributions of the comparison vectors. It iteratively directly estimates the density ratio between the matches and the non-matches, using the Kullback-Leibler Importance Estimation Procedure (KLIEP). For details see Sugiyama et al. (2008).

The mec() function allows the construction of the predicted set of matches using its estimated size or the bisection procedure, described in Lee et al. (2022), based on a target false link rate (FLR) or missing match rate (MMR). To use the second option, set set_construction = "flr" or set_construction = "mmr" and specify a target error rate using the target_rate argument.

The assumption that A and B contain no duplicate records might be relaxed by allowing A to have duplicates. To do so, set duplicates_in_A = TRUE.

Value

Returns a list containing:

• M_est – a data.table with predicted matches,

• n_M_est – estimated classification set size,

• flr_est – estimated false link rate (FLR),

• mmr_est – estimated missing match rate (MMR),

• iter_bisection – the number of iterations in the bisection procedure,

• b_vars – a character vector of variables used for the "binary" method (with the prefix "gamma_"),

• cpar_vars – a character vector of variables used for the "continuous_parametric" method (with the prefix "gamma_"),

• cnonpar_vars – a character vector of variables used for the "continuous_nonparametric" method (with the prefix "gamma_"),

• hm_vars – a character vector of variables used for the "hit_miss" method (with the prefix "gamma_"),

• b_params – parameters estimated using the "binary" method,

• cpar_params – parameters estimated using the "continuous_parametric" method,

• hm_params – parameters estimated using the "hit_miss" method,

• ratio_kliep – a result of the kliep() function,

• variables – a character vector of key variables used for comparison,

• set_construction – a method for constructing the predicted set of matches,

• eval_metrics – standard metrics for quality assessment (if true_matches is provided),

• confusion – confusion matrix (if true_matches is provided).

Author(s)

Adam Struzik

References

Lee, D., Zhang, L.-C. and Kim, J. K. (2022). Maximum entropy classification for record linkage. Survey Methodology, Statistics Canada, Catalogue No. 12-001-X, Vol. 48, No. 1.

Vo, T. H., Chauvet, G., Happe, A., Oger, E., Paquelet, S., and Garès, V. (2023). Extending the Fellegi-Sunter record linkage model for mixed-type data with application to the French national health data system. Computational Statistics & Data Analysis, 179, 107656.

Sugiyama, M., Suzuki, T., Nakajima, S. et al. Direct importance estimation for covariate shift adaptation. Ann Inst Stat Math 60, 699–746 (2008). doi:10.1007/s10463-008-0197-x

Examples

df_1 <- data.frame(
  name = c("Emma", "Liam", "Olivia", "Noah", "Ava",
           "Ethan", "Sophia", "Mason", "Isabella", "James"),
  surname = c("Smith", "Johnson", "Williams", "Brown", "Jones",
              "Garcia", "Miller", "Davis", "Rodriguez", "Wilson"),
  city = c("New York", "Los Angeles", "Chicago", "Houston", "Phoenix",
           "Philadelphia", "San Antonio", "San Diego", "Dallas", "San Jose")
)

df_2 <- data.frame(
  name = c(
    "Emma", "Liam", "Olivia", "Noah",
    "Ava", "Ehtan", "Sopia", "Mson",
    "Charlotte", "Benjamin", "Amelia", "Lucas"
  ),
  surname = c(
     "Smith", "Johnson", "Williams", "Brown",
    "Jnes", "Garca", "Miler", "Dvis",
    "Martinez", "Lee", "Hernandez", "Clark"
  ),
  city = c(
    "New York", "Los Angeles", "Chicago", "Houston",
    "Phonix", "Philadelpia", "San Antnio", "San Dieg",
    "Seattle", "Miami", "Boston", "Denver"
  )
)
true_matches <- data.frame(
  "a" = 1:8,
  "b" = 1:8
)

variables <- c("name", "surname", "city")
comparators <- list(
  "name" = jarowinkler_complement(),
  "surname" = jarowinkler_complement(),
  "city" = jarowinkler_complement()
)
methods <- list(
  "name" = "continuous_parametric",
  "surname" = "continuous_parametric",
  "city" = "continuous_parametric"
)

set.seed(1)
result <- mec(A = df_1, B = df_2,
              variables = variables,
              comparators = comparators,
              methods = methods,
              true_matches = true_matches)
result

Blocked Unsupervised Maximum Entropy Classifier for Record Linkage

Description

Runs graph-based blocking using blocking(), fits one pooled unsupervised maximum entropy classifier (MEC) on selected within-block pairs, and applies the fitted density-ratio model blockwise.

Usage

mec_blocking(
  A,
  B,
  variables,
  comparators = NULL,
  methods = NULL,
  blocking_x = NULL,
  blocking_y = NULL,
  blocking_variables = variables,
  blocking_sep = " ",
  controls_blocking = list(),
  min_training_pairs = NULL,
  min_training_nonmatches = NULL,
  block_sampling_seed = NULL,
  nonmatch_sample_size = NULL,
  nonmatch_sampling_seed = NULL,
  prob_ratio = "2",
  start_params = NULL,
  nonpar_hurdle = TRUE,
  fixed_method = "Newton",
  delta = 0.5,
  eps = 0.05,
  max_iter_em = 10,
  tol_em = 1,
  controls_nleqslv = list(),
  controls_kliep = control_kliep(),
  true_matches = NULL,
  keep_blocking_result = FALSE,
  keep_training_data = FALSE,
  verbose = FALSE
)

Arguments

Details

The function assumes one-to-one linkage. The blocking stage defines disjoint bipartite blocks. MEC is trained once on the pooled union of selected within-block Cartesian products and is then applied separately in each final block. The ANN distance returned by blocking() is not used.

If both min_training_pairs and min_training_nonmatches are NULL, all final blocks are used for pooled training. If both are supplied, blocks are sampled without replacement until both thresholds are reached. Supplying only one threshold is an error.

Nonmatch distribution parameters are estimated from a simple random sample from the full Cartesian product of A and B, not from the selected training blocks.

Value

Returns a list of class "mec_blocking" containing:

• M_est – a data.table with predicted matches and columns a, b, block, and ratio,

• n_M_est – estimated total number of matches across all blocks,

• flr_est – estimated false link rate (FLR),

• mmr_est – estimated missing match rate (MMR),

• training_rule – training-block selection rule used by the function,

• block_estimates – a data.table with block-level match-count and error-rate estimates,

• training_blocks – a data.table with blocks selected for pooled MEC training,

• block_summary – a data.table describing the final disjoint blocks,

• excluded_records – a list with records from A and B excluded by blocking,

• pooled_model – fitted pooled MEC density-ratio model used for blockwise scoring,

• b_vars – variables used for the "binary" method, with the prefix "gamma_",

• cpar_vars – variables used for the "continuous_parametric" method, with the prefix "gamma_",

• cnonpar_vars – variables used for the "continuous_nonparametric" method, currently NULL,

• hm_vars – variables used for the "hit_miss" method, currently NULL,

• b_params – parameters estimated using the "binary" method,

• cpar_params – parameters estimated using the "continuous_parametric" method,

• cnonpar_params – parameters estimated using the "continuous_nonparametric" method, currently NULL,

• hm_params – parameters estimated using the "hit_miss" method, currently NULL,

• ratio_kliep – result of kliep(), currently NULL,

• ratio_kliep_list – variable-specific KLIEP results, currently NULL,

• variables – key variables used for comparison,

• comparators – comparison functions used to create comparison vectors,

• methods – MEC estimation methods used for the key variables,

• nonmatch_sample_size – number of full Cartesian-product pairs used to estimate nonmatch parameters,

• nonmatch_sampling_seed – seed used for nonmatch-pair sampling,

• prob_ratio – probability/density ratio type used for blockwise match-count estimation,

• delta – tolerance for changes in the estimated number of matches,

• eps – tolerance for changes in model parameters,

• controls_nleqslv – controls passed to nleqslv(),

• controls_blocking – additional arguments passed to blocking(),

• blocking_result – raw object returned by blocking() if keep_blocking_result = TRUE; otherwise NULL,

• training_Omega – pooled training comparison vectors if keep_training_data = TRUE; otherwise NULL,

• blocking_eval – blocking diagnostics if true_matches is provided; otherwise NULL,

• eval_metrics – standard linkage quality metrics if true_matches is provided; otherwise NULL,

• confusion – confusion matrix if true_matches is provided; otherwise NULL.

Author(s)

Adam Struzik

Examples

df_1 <- data.frame(
  name = c("Emma", "Liam", "Olivia", "Noah", "Ava"),
  surname = c("Smith", "Jones", "Brown", "Davis", "Miller"),
  city = c("Boston", "Boston", "Austin", "Austin", "Denver")
)
df_2 <- data.frame(
  name = c("Emma", "Liam", "Olivia", "Noah", "Ava"),
  surname = c("Smith", "Jones", "Brown", "Davis", "Miller"),
  city = c("Boston", "Boston", "Austin", "Austin", "Denver")
)

blocking_x <- matrix(
  c(1, 0, 0, 1, 1, 1, 2, 0, 0, 2),
  ncol = 2,
  byrow = TRUE
)
blocking_y <- blocking_x

result <- mec_blocking(
  A = df_1,
  B = df_2,
  variables = c("name", "surname", "city"),
  blocking_x = blocking_x,
  blocking_y = blocking_y,
  controls_blocking = list(
    representation = "custom_matrix",
    ann = "kd",
    distance = "euclidean",
    seed = 1
  ),
  true_matches = data.frame(a = 1:5, b = 1:5)
)
result

Predict Matches Based on a Given Record Linkage Model

Description

Predicts matches between records in two datasets based on a given record linkage model, using the maximum entropy classification (MEC) algorithm (see Lee et al. (2022)).

Usage

## S3 method for class 'rec_lin_model'
predict(
  object,
  newdata_A,
  newdata_B,
  duplicates_in_A = FALSE,
  set_construction = c("size", "flr", "mmr"),
  fixed_method = "Newton",
  target_rate = 0.03,
  tol = 0.005,
  max_iter = 50,
  data_type = c("data.frame", "data.table", "matrix"),
  true_matches = NULL,
  verbose = FALSE,
  ...
)

Arguments

Details

The predict.rec_lin_model() method estimates the probability/density ratio between matches and non-matches for pairs in given datasets, based on a model obtained using the train_rec_lin() or custom_rec_lin_model(). Then, it estimates the number of matches and returns the predicted matches, using the maximum entropy classification (MEC) algorithm (see Lee et al. (2022)).

The predict.rec_lin_model() method allows the construction of the predicted set of matches using its estimated size or the bisection procedure, described in Lee et al. (2022), based on a target false link rate (FLR) or missing match rate (MMR). To use the second option, set set_construction = "flr" or set_construction = "mmr" and specify a target error rate using the target_rate argument.

By default, the function assumes that the datasets newdata_A and newdata_B contain no duplicate records. This assumption might be relaxed by allowing newdata_A to have duplicates. To do so, set duplicates_in_A = TRUE.

Value

Returns a list containing:

• M_est – a data.table with predicted matches,

• set_construction – a method for constructing the predicted set of matches,

• n_M_est – estimated classification set size,

• flr_est – estimated false link rate (FLR),

• mmr_est – estimated missing match rate (MMR),

• iter – the number of iterations in the bisection procedure,

• eval_metrics – standard metrics for quality assessment, if true_matches is provided,

• confusion – confusion matrix, if true_matches is provided.

Author(s)

Adam Struzik

References

Lee, D., Zhang, L.-C. and Kim, J. K. (2022). Maximum entropy classification for record linkage. Survey Methodology, Statistics Canada, Catalogue No. 12-001-X, Vol. 48, No. 1.

Vo, T. H., Chauvet, G., Happe, A., Oger, E., Paquelet, S., and Garès, V. (2023). Extending the Fellegi-Sunter record linkage model for mixed-type data with application to the French national health data system. Computational Statistics & Data Analysis, 179, 107656.

Sugiyama, M., Suzuki, T., Nakajima, S. et al. Direct importance estimation for covariate shift adaptation. Ann Inst Stat Math 60, 699–746 (2008). doi:10.1007/s10463-008-0197-x

Examples

df_1 <- data.frame(
  "name" = c("James", "Emma", "William", "Olivia", "Thomas",
  "Sophie", "Harry", "Amelia", "George", "Isabella"),
  "surname" = c("Smith", "Johnson", "Brown", "Taylor", "Wilson",
  "Davis", "Clark", "Harris", "Lewis", "Walker")
)
 df_2 <- data.frame(
  "name" = c("James", "Ema", "Wimliam", "Olivia", "Charlotte",
  "Henry", "Lucy", "Edward", "Alice", "Jack"),
  "surname" = c("Smith", "Johnson", "Bron", "Tailor", "Moore",
  "Evans", "Hall", "Wright", "Green", "King")
)
comparators <- list("name" = jarowinkler_complement(),
                    "surname" = jarowinkler_complement())
matches <- data.frame("a" = 1:4, "b" = 1:4)
methods <- list("name" = "continuous_nonparametric",
                "surname" = "continuous_nonparametric")
model <- train_rec_lin(A = df_1, B = df_2, matches = matches,
                       variables = c("name", "surname"),
                       comparators = comparators,
                       methods = methods)

df_new_1 <- data.frame(
  "name" = c("John", "Emily", "Mark", "Anna", "David"),
  "surname" = c("Smith", "Johnson", "Taylor", "Williams", "Brown")
)
df_new_2 <- data.frame(
  "name" = c("John", "Emely", "Mark", "Michael"),
  "surname" = c("Smitth", "Johnson", "Tailor", "Henders")
)
predict(model, df_new_1, df_new_2)

Train a Record Linkage Model

Description

Trains a supervised record linkage model using probability or density ratio estimation, based on Lee et al. (2022), with several extensions.

Usage

train_rec_lin(
  A,
  B,
  matches,
  variables,
  comparators = NULL,
  methods = NULL,
  prob_ratio = NULL,
  nonpar_hurdle = TRUE,
  controls_nleqslv = list(),
  controls_kliep = control_kliep(),
  verbose = FALSE
)

Arguments

Details

Consider two datasets: A and B. Let the bipartite comparison space \Omega = A \times B consist of matches M and non-matches U between the records in files A and B. For any pair of records (a,b) \in \Omega, let \pmb{\gamma}_{ab} = (\gamma_{ab}^1,\gamma_{ab}^2, \ldots,\gamma_{ab}^K)' be the comparison vector between a set of key variables. The original MEC algorithm uses the binary comparison function to evaluate record pairs across two datasets. However, this approach may be insufficient when handling datasets with frequent errors across multiple variables.

We propose the use of continuous comparison functions to address the limitations of binary comparison methods. We consider every semi-metric, i.e., a function d: A \times B \to \mathbb{R}, satisfying the following conditions:

1. d(x,y) \geq 0,

2. d(x,y) = 0 if and only if x = y,

3. d(x,y) = d(y,x).

For example, we can use the Jaro-Winkler distance for character variables (which is implemented in the automatedRecLin package as jarowinkler_complement()) or the Euclidean distance for numerical variables. The automatedRecLin package allows the use of a different comparison function for each key variable (which should be specified as a list in the comparators argument). The default function for each key variable is cmp_identical() (the binary comparison function).

The train_rec_lin() function is used to train a record linkage model, when M and U are known (which might later serve as a classifier for pairs outside \Omega). It offers different approaches to estimate the probability/density ratio between matches and non-matches, which should be specified as a list in the methods argument. The method suitable for the binary comparison function is "binary", which is also the default method for each variable.

For the continuous semi-metrics we suggest the usage of "continuous_parametric" or "continuous_nonparametric" method. The "continuous_parametric" method assumes that \gamma_{ab}^k|M and \gamma_{ab}^k|U follow hurdle Gamma distributions. The density function of a hurdle Gamma distribution is characterized by three parameters p_0 \in (0,1) and \alpha, \beta > 0 as follows:

f(x;p_0,\alpha,\beta) = p_0^{\mathbb{I}(x = 0)}[(1 - p_0) v(x;\alpha,\beta)]^{\mathbb{I}(x > 0)},

where

v(x;\alpha,\beta) = \frac{\beta^{\alpha} x^{\alpha - 1} \exp(-\beta x)} {\Gamma(\alpha)}

is the density function of a Gamma distribution (for details see Vo et al. (2023)). The "continuous_nonparametric" method does not assume anything about the distributions of the comparison vectors. It directly estimates the density ratio between the matches and the non-matches, using the Kullback-Leibler Importance Estimation Procedure (KLIEP). For details see Sugiyama et al. (2008).

Value

Returns a list containing:

• b_vars – a character vector of variables used for the "binary" method (with the prefix "gamma_"),

• cpar_vars – a character vector of variables used for the "continuous_parametric" method (with the prefix "gamma_"),

• cnonpar_vars – a character vector of variables used for the "continuous_nonparametric" method (with the prefix "gamma_"),

• b_params – parameters estimated using the "binary" method,

• cpar_params – parameters estimated using the "continuous_parametric" method,

• cnonpar_params – probability of exact matching estimated using the "continuous_nonparametric" method,

• ratio_kliep – a result of the kliep() function,

• ratio_kliep_list – an object containing the results of the kliep() function,

• ml_model – here NULL,

• pi_est – a prior probability of matching,

• match_prop – proportion of matches in the smaller dataset,

• variables – a character vector of key variables used for comparison,

• comparators – a list of functions used to compare pairs of records,

• methods – a list of methods used for estimation,

• "prob_ratio" – probability/density ratio type.

Author(s)

Adam Struzik

References

Lee, D., Zhang, L.-C. and Kim, J. K. (2022). Maximum entropy classification for record linkage. Survey Methodology, Statistics Canada, Catalogue No. 12-001-X, Vol. 48, No. 1.

Vo, T. H., Chauvet, G., Happe, A., Oger, E., Paquelet, S., and Garès, V. (2023). Extending the Fellegi-Sunter record linkage model for mixed-type data with application to the French national health data system. Computational Statistics & Data Analysis, 179, 107656.

Sugiyama, M., Suzuki, T., Nakajima, S. et al. Direct importance estimation for covariate shift adaptation. Ann Inst Stat Math 60, 699–746 (2008). doi:10.1007/s10463-008-0197-x

Examples

df_1 <- data.frame(
  "name" = c("James", "Emma", "William", "Olivia", "Thomas",
  "Sophie", "Harry", "Amelia", "George", "Isabella"),
  "surname" = c("Smith", "Johnson", "Brown", "Taylor", "Wilson",
  "Davis", "Clark", "Harris", "Lewis", "Walker")
)
 df_2 <- data.frame(
  "name" = c("James", "Ema", "Wimliam", "Olivia", "Charlotte",
  "Henry", "Lucy", "Edward", "Alice", "Jack"),
  "surname" = c("Smith", "Johnson", "Bron", "Tailor", "Moore",
  "Evans", "Hall", "Wright", "Green", "King")
)
comparators <- list("name" = jarowinkler_complement(),
                    "surname" = jarowinkler_complement())
matches <- data.frame("a" = 1:4, "b" = 1:4)
methods <- list("name" = "continuous_nonparametric",
                "surname" = "continuous_nonparametric")
model <- train_rec_lin(A = df_1, B = df_2, matches = matches,
                       variables = c("name", "surname"),
                       comparators = comparators,
                       methods = methods)
model