Tutorial

In this tutorial, we will show you how to run LiNGAM algorithms and see the results. We will also show you how to run the bootstrap method and check the results.

The following packages must be installed in order to run this tutorial. And import if necessary:

• numpy

• pandas

• scikit-learn

• graphviz

• statsmodels

Contents:

• DirectLiNGAM
  • Model
  • Import and settings
  • Test data
  • Causal Discovery
  • Independence between error variables
• Bootstrap
  • Import and settings
  • Test data
  • Bootstrapping
  • Causal Directions
  • Directed Acyclic Graphs
  • Probability
  • Total Causal Effects
  • Bootstrap Probability of Path
• How to use prior knowledge in DirectLiNGAM
  • Import and settings
  • Utility function
  • Test data
  • Make Prior Knowledge Matrix
  • Causal Discovery
• MultiGroupDirectLiNGAM
  • Model
  • Import and settings
  • Test data
  • Causal Discovery
  • Independence between error variables
  • Bootstrapping
  • Causal Directions
  • Directed Acyclic Graphs
  • Probability
  • Total Causal Effects
  • Bootstrap Probability of Path
• Total Effect
• Causal Effect on predicted variables
  • Identification of Feature with Greatest Causal Influence on Prediction
  • Estimation of Optimal Intervention
  • Use a known causal model
• VARLiNGAM
  • Model
  • Import and settings
  • Test data
  • Causal Discovery
  • Independence between error variables
  • Bootstrap
  • Causal Directions
  • Directed Acyclic Graphs
  • Probability
  • Total Causal Effects
• VARMALiNGAM
  • Model
  • Import and settings
  • Test data
  • Causal Discovery
  • Independence between error variables
  • Bootstrap
  • Causal Directions
  • Directed Acyclic Graphs
  • Probability
  • Total Causal Effects
• Longitudinal LiNGAM
  • Model
  • Import and settings
  • Test data
  • Causal Discovery
  • Independence between error variables
  • Bootstrapping
  • Causal Directions
  • Directed Acyclic Graphs
  • Probability
  • Total Causal Effects
• BottomUpParceLiNGAM
  • Model
  • Import and settings
  • Test data
  • Causal Discovery
  • Independence between error variables
  • Bootstrapping
  • Causal Directions
  • Directed Acyclic Graphs
  • Probability
  • Total Causal Effects
  • Bootstrap Probability of Path
• How to use prior knowledge in BottomUpParceLiNGAM
  • Import and settings
  • Utility function
  • Test data
  • Make Prior Knowledge Matrix
  • Causal Discovery
• RCD
  • Model
  • Import and settings
  • Test data
  • Causal Discovery
  • Independence between error variables
  • Bootstrapping
  • Causal Directions
  • Directed Acyclic Graphs
  • Probability
  • Total Causal Effects
  • Bootstrap Probability of Path
• Draw Causal Graph
  • Import and settings
  • Draw the result of LiNGAM
  • Save graph
  • Draw the result of LiNGAM with prediction model
  • Highlight paths between specified nodes
  • Draw the result of LiNGAM with emphasis on descendants and ancestors
  • Draw the result of Bootstrap with emphasis on descendants and ancestors
• LiNA
  • Model
  • Import and settings
  • Single-domain test data
  • Causal Discovery for single-domain data
  • Multi-domain test data
  • Causal Discovery for multi-domain data
• RESIT
  • Model
  • Import and settings
  • Test data
  • Causal Discovery
  • Bootstrapping
  • Causal Directions
  • Directed Acyclic Graphs
  • Probability
  • Bootstrap Probability of Path
• LiM
  • Model
  • Import and settings
  • Test data
  • Causal Discovery for linear mixed data
• CAM-UV
  • Model
  • Import and settings
  • Test data
  • Causal Discovery
• MultiGroupRCD
  • Import and settings
  • Test data
  • Causal Discovery
• Finding ancestors of each variable
  • Import and settings
  • Test data
  • Extract the ancestors of each observed variable
• F-correlation
  • Import and settings
  • Test data
  • Linear correlated data (Uncorrelated)
  • Linear correlated data (Strongly correlated)
  • Non-linear correlated data (Quadratic function)
  • Non-linear correlated data (Sin function)
• Visualization of nonlinear causal effect
  • Import and settings
  • Data generation
  • Causal discovery
  • Visualization
• EvaluateModelFit
  • Import and settings
  • When all variables are continuous data
  • Test data
  • Causal Discovery
  • Evaluation
  • When the data has hidden common causes
  • Test data
  • Causal Discovery
  • When the data has ordinal variables
  • Test data
• Bootstrap with imputation
  • Import and settings
  • Test data
  • Causal discovery with missing data
  • Checking results
  • Getting bootstrap samples and imputed bootstrap samples
• HighDimDirectLiNGAM
  • Import and settings
  • Test data
  • Causal Discovery