REMFlow documentation

REMFlow fits relational event models to ordered sender-receiver data. It constructs event histories and risk sets, calculates time-varying network statistics, and estimates ordinal, exact-time, actor-oriented, and duration models. The fitted objects provide coefficient estimates, event probabilities, residuals, ranks, and other model diagnostics.

Installation

Install the released remflow package from PyPI:

python -m pip install remflow

A source checkout is required only for development. See the Quick start page for optional JAX installation and development setup.

Getting started

• Introduction
  • Implementation choices
• Quick start
  • Install
  • Fit a model
  • Select a JAX device
  • Where to go next
• Usage guide
  • Input data
  • Build an event history
  • Calculate statistics
  • Fit the model
  • Actor-oriented models
  • Duration events
  • GPU backend
  • High-level relational event model
  • Misinformation analysis
  • Support notes

Studies and performance

• Performance evaluation
  • Reference workloads
  • What a result records
  • Reading the numbers
  • Running the benchmarks
• Classroom relational-event study
  • Analysis reproduced here
  • What the comparison means
  • Run the example
  • Data files

Reference

• API reference
  • Typical workflow
  • Event histories and risk sets
  • Statistics and formulas
  • Estimation and results
  • Backend selection
  • High-level models

Indices and tables

• Index

• Module Index

• Search Page