Python Record Linkage Toolkit Documentation

All you need to start linking records.

First steps

• About
  • Introduction
  • What is record linkage?
  • How to link records?
• Installation
  • Python version support
  • Installation
  • Dependencies
• Link two datasets
  • Introduction
  • Make record pairs
  • Compare records
  • Full code
• Data deduplication
  • Introduction
  • Make record pairs
  • Compare records
  • Full code

Record linkage

• 0. Preprocessing
  • Cleaning
  • Phonetic encoding
• 1. Indexing
  • recordlinkage.Index object
  • Algorithms
  • User-defined algorithms
  • Examples
• 2. Comparing
  • recordlinkage.Compare object
  • Algorithms
  • User-defined algorithms
  • Examples
• 3. Classification
  • Classifiers
  • Adapters
  • User-defined algorithms
  • Examples
  • Network
• 4. Evaluation
  • reduction_ratio()
  • true_positives()
  • true_negatives()
  • false_positives()
  • false_negatives()
  • confusion_matrix()
  • precision()
  • recall()
  • accuracy()
  • specificity()
  • fscore()
  • max_pairs()
  • full_index_size()
• Datasets
  • load_krebsregister()
  • load_febrl1()
  • load_febrl2()
  • load_febrl3()
  • load_febrl4()
  • binary_vectors()
• Miscellaneous
  • index_split()
  • get_option()
  • set_option()
  • reset_option()
  • describe_option()

Miscellaneous

• Annotation
  • Generate annotation file
  • Manual labeling
  • Export/read annotation file
• Classification algorithms
  • Supervised learning
  • Unsupervised learning
• Performance
  • Indexing
  • Comparing

Developers

• Contributing
• Release notes