Skip to content

on_boarding.md

Latest commit

 

History

History
84 lines (53 loc) · 7.29 KB

on_boarding.md

File metadata and controls

84 lines (53 loc) · 7.29 KB
graph LR
    Data_Management["Data Management"]
    Graph_Representation_Layer["Graph Representation Layer"]
    Causal_Discovery_Algorithms["Causal Discovery Algorithms"]
    Statistical_Test_Engine["Statistical Test Engine"]
    Auxiliary_Utilities["Auxiliary Utilities"]
    Data_Management -- "Provides preprocessed input data" --> Causal_Discovery_Algorithms
    Causal_Discovery_Algorithms -- "Queries for conditional independence relationships" --> Statistical_Test_Engine
    Causal_Discovery_Algorithms -- "Constructs and updates causal graph structures" --> Graph_Representation_Layer
    Causal_Discovery_Algorithms -- "Outputs the final inferred causal graph" --> Graph_Representation_Layer
    Auxiliary_Utilities -- "Performs transformations and manipulations on graph structures" --> Graph_Representation_Layer
    click Data_Management href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main/causal-learn/Data_Management.md" "Details"
    click Graph_Representation_Layer href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main/causal-learn/Graph_Representation_Layer.md" "Details"
    click Causal_Discovery_Algorithms href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main/causal-learn/Causal_Discovery_Algorithms.md" "Details"
    click Statistical_Test_Engine href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main/causal-learn/Statistical_Test_Engine.md" "Details"
Loading

CodeBoardingDemoContact

Details

The causal-learn project is structured around a core set of components designed for comprehensive causal discovery. The process begins with the Data Management component, responsible for ingesting and preparing raw data. This preprocessed data then feeds into the Causal Discovery Algorithms, which represent the analytical heart of the system. These algorithms leverage the Statistical Test Engine to perform crucial conditional independence tests and other statistical evaluations, guiding the discovery process. All discovered causal relationships and graph structures are managed and represented by the Graph Representation Layer. Finally, the Auxiliary Utilities provide a suite of supporting functions for graph manipulation, conversions, and other common tasks, ensuring flexibility and extensibility within the causal inference workflow. This modular design facilitates clear data flow, allowing for independent development and integration of new algorithms and statistical tests.

Data Management [Expand]

Handles the loading, validation, and initial preparation of input data for all causal discovery algorithms. It serves as the primary data ingress point.

Related Classes/Methods:

Graph Representation Layer [Expand]

Provides the foundational data structures and interfaces for defining, storing, and managing various types of causal graphs (e.g., DAGs, PAGs, general graphs), including their nodes and edges.

Related Classes/Methods:

Causal Discovery Algorithms [Expand]

The central component encapsulating all implemented causal discovery methodologies, including constraint-based, FCM-based, permutation-based, hidden causal representation learning, and score-based algorithms. It orchestrates the discovery process.

Related Classes/Methods:

Statistical Test Engine [Expand]

Offers a unified interface for various statistical tests, including conditional independence tests (CIT) and non-linear independence criteria (HSIC), crucial for many causal discovery algorithms.

Related Classes/Methods:

Auxiliary Utilities

A collection of supporting functions for graph manipulation, conversion utilities (e.g., DAG to PAG), PC/FCI-specific helpers, statistical resampling (bootstrapping), and causal effect estimation.

Related Classes/Methods:

FAQ