Package index • irelink

Linkage Specification

Define the comparisons and blocking rules that drive the model.

il_spec(): Create an Empty Linkage Specification
il_compare(): Add a Comparison Layer to a Specification
il_block_on(): Add a Prediction Blocking Rule
block_on(): Create a Training-Time Blocking Rule
il_transform(): Compose Multiple Transforms into a Chain
il_substr(): Extract a Substring Column Transform
il_regex_extract(): Regex Extraction Column Transform
il_nullif(): Replace a Value with NA Column Transform
il_cast_to_string(): Cast to String Column Transform
il_try_parse_date(): Try-Parse Date Column Transform
il_try_parse_timestamp(): Try-Parse Timestamp Column Transform
il_array_element(): Array Element Column Transform
is_il_spec(): Test if an Object is an irelink Specification

Comparison Levels

Building blocks for scoring how similar two records are on a given field.

cl_exact(): Exact Equality Comparison
cl_levenshtein(): Levenshtein Edit-Distance Comparison
cl_damerau_levenshtein(): Damerau-Levenshtein Edit-Distance Comparison
cl_jaro(): Jaro String Similarity Comparison
cl_jaro_winkler(): Jaro-Winkler String Similarity Comparison
cl_jaccard(): Jaccard Set Similarity Comparison
cl_cosine(): Cosine Similarity Comparison
cl_numeric_diff(): Numeric Absolute Difference Comparison
cl_pct_diff(): Numeric Percentage Difference Comparison
cl_date_diff(): Date Difference Comparison
cl_time_diff(): Time Difference Comparison
cl_geo_distance(): Geographic Distance Comparison
cl_array_intersect(): Array Intersection Comparison
cl_array_subset(): Array Subset Comparison
cl_array_min_distance(): Pairwise Array Minimum Distance Comparison
cl_columns_reversed(): Swap Detection for Two Columns
cl_custom(): Custom SQL Comparison
cl_literal(): Literal Value Comparison
cl_null(): Null / Missing Value Level
cl_else(): Catch-All Else Level
cl_levels(): Compose Custom Comparison Levels
cl_and(): Combine Comparison Conditions with AND
cl_or(): Combine Comparison Conditions with OR
cl_not(): Negate a Comparison Condition

Phonetic Transforms

Phonetic encoding functions for blocking and comparison on name fields.

il_soundex() il_metaphone() il_dmetaphone(): Phonetic Transform Functions
cl_soundex(): Soundex Phonetic Comparison

Domain-Specific Comparisons

Pre-configured multi-level comparisons for common field types.

cl_name(): Personal Name Comparison
cl_first_last_name(): First Name and Last Name Comparison with Swap Detection
cl_forename_surname(): Forename and Surname Comparison with Swap Detection
cl_dob(): Date of Birth Comparison
cl_email(): Email Address Comparison
cl_zip_code(): ZIP Code Comparison
cl_postcode(): Postcode Comparison

Model Fitting

Create and train a probabilistic linkage model.

il_model(): Create a Linkage Model
il_estimate_u(): Estimate Non-Match (u) Parameters
il_estimate_em(): Train Parameters via Expectation-Maximization
il_estimate_prior(): Estimate the Prior Match Probability
il_prior_prevalence(): Add a Prevalence Prior
il_prior_m(): Add a Matched-Class Comparison Prior
il_constrain_m(): Add a Fixed Matched-Class Constraint
il_estimate_m_from_column(): Estimate Match (m) Parameters from a Label Column
il_estimate_m_from_labels(): Estimate Match (m) Parameters from Labeled Data
is_il_model(): Test if an Object is an irelink Model

Prediction and Clustering

Score record pairs and resolve them into linked entities.

predict(<il_model>): Score Record Pairs from a Trained Model
il_cluster(): Cluster Scored Pairs into Entities
il_deterministic_link(): Deterministic Record Linkage
il_find_matches(): Find Matches for New Records
il_score_patterns(): Score Comparison Patterns
il_score_missing_edges(): Score Missing Edges Within Clusters

Model Inspection

Examine parameters, weights, and training diagnostics.

il_parameters(): Extract Model Parameters
il_priors(): Inspect Model Priors
il_constraints(): Inspect Model Constraints
il_weights(): Extract Match Weights by Comparison Level
il_training_history(): Extract EM Training History
autoplot(<il_training_history>): Plot EM Training History
il_waterfall(): Extract Waterfall Data for a Single Pair
il_compare_records(): Compare Two Individual Records
il_string_similarity(): Compute String Similarity Scores
autoplot(<il_string_similarity>): Comparator Score Bar Chart
il_tf_chart(): Term Frequency Adjustment Chart
il_comparison_vectors(): Comparison Vector Distribution
autoplot(<il_comparison_vectors>): Plot Comparison Vector Distribution
print(<il_model>): Print an irelink Model
print(<il_spec>): Print an irelink Specification
summary(<il_model>): Summarize an irelink Model
autoplot(<il_model>): Quick Match-Weights Plot for a Model
autoplot(<il_compared>): Quick Plot for Scored Pairs

Evaluation

Assess model quality against labeled data.

il_accuracy(): Accuracy Metrics Across Thresholds
il_confusion_matrix(): Confusion Matrix at a Threshold
il_cluster_confusion_matrix(): Cluster-Level Confusion Matrix for Deduplication
labels_from_column(): Derive Pairwise Labels from a Ground-Truth Column
il_precision_recall(): Compute Precision-Recall Curve Data
il_roc(): Compute ROC Curve Data
il_errors(): Identify Prediction Errors
il_unlinkables(): Compute Unlinkable Records
il_graph_metrics(): Compute Graph Metrics for Clusters
autoplot(<il_accuracy>): Plot Accuracy Metrics Across Thresholds
autoplot(<il_roc>): Plot ROC Curve
autoplot(<il_precision_recall>): Plot Precision–Recall Curve
autoplot(<il_unlinkables>): Plot Unlinkables Curve

Blocking

Suggest and evaluate blocking rules to reduce candidate pairs.

il_suggest_blocking(): Suggest Blocking Rules
il_find_blocking_below(): Find Blocking Rules Below a Pair-Count Threshold
block_from_labels(): Derive Blocking Rules from Labeled Pairs

Data Profiling

Explore and summarize input data before linkage.

il_completeness(): Column Completeness Across Datasets
autoplot(<il_completeness>): Plot Column Completeness
il_count_pairs(): Count Candidate Pairs Under Blocking Rules
autoplot(<il_count_pairs>): Plot Blocking Rule Pair Counts
il_largest_blocks(): Identify the Largest Blocking Bins
il_profile(): Profile Column Value Distributions
autoplot(<il_profile>): Plot Column Value Profiles
il_comparator_score(): Batch String Similarity Scores
autoplot(<il_comparator_score>): Plot Batch Comparator Scores
il_comparator_threshold_chart(): Comparator Score Threshold Chart
il_phonetic_chart(): Phonetic Match Chart
il_register_tf(): Register Pre-Computed Term Frequency Tables

Persistence and Utilities

Save, load, and manage linkage models and resources.

il_save(): Save a Model to Disk
il_load(): Load a Saved Model
il_attach(): Attach a Saved Model to Fresh Data
il_cleanup(): Remove Model-Owned Temporary Tables from Database
il_cleanup_all(): Remove All irelink Temporary Tables from a Database
irelink irelink-package: irelink: Fast Probabilistic Record Linkage

Datasets

Bundled benchmark datasets from the splink ecosystem.

fake_20: Fake 20: Minimal Deduplication Example
fake_1000: Splink Fake 1000: Deduplication Benchmark
fake_1000_labels: Splink Fake 1000: Clerical Pairwise Labels
febrl4a: FEBRL 4a: Record Linkage Original Records
febrl4b: FEBRL 4b: Record Linkage Duplicate Records

Unit Helpers

Lightweight constructors for physical and temporal units.

days(): Create a Duration in Days
months(): Create a Duration in Months
years(): Create a Duration in Years
hours(): Create a Duration in Hours
minutes(): Create a Duration in Minutes
seconds(): Create a Duration in Seconds
km(): Create a Distance in Kilometres
mi(): Create a Distance in Miles