Welcome to hlink’s documentation!

• Introduction
  • Overview
• Installation
  • Requirements
  • Installing from PyPI
  • Installing from source
• Link Tasks
  • Preprocessing
  • Training and Household Training
  • Matching
  • Household Matching
  • Model Exploration and Household Model Exploration
  • Reporting
• Running hlink
  • Using hlink as a Library
  • Interactive Mode
• Advanced Workflows
  • Export training data after generating features to reuse in different linking years
  • ML model exploration and export of lists of potential false positives/negatives in training data
• Configuration
  • Basic Config File
  • Advanced Config File
  • Top level configs
  • Data sources
  • Filter
  • Column Mappings
  • Substitution Columns
  • Feature Selections
  • Potential Matches Universe
  • Blocking
  • Comparisons
  • Household Comparisons
  • Comparison Features
  • Pipeline-generated Features
  • Training and models
  • Household training and models

Configuration API

Configuration API

• Column Mappings
  • Basic Usage
  • Advanced Usage
  • Transforms
• Comparisons
  • Overview
  • Comparison Types
  • Defining Multiple Comparisons
• Comparison Features
  • Comparison Types
  • Feature add-ons
  • Aggregate Features
  • Household Aggregate Features
• Feature Selection
  • bigrams
  • sql_condition
  • array
  • union
  • soundex
  • power
• Pipeline Features
  • Transformer types
• Substitutions
  • 1:1 substitution by data table
  • Substitution by regex word replace
• Models
  • random_forest
  • probit
  • logistic_regression
  • decision_tree
  • gradient_boosted_trees