FCA Algorithms in Python

Python implementation of algorithms from Formal Concept Analysis (FCA), including concept lattice construction, attribute reduction, reduct computation and implication analysis.

Overview

The workflow is:

1. Define an object–attribute relation in ConExp.
2. Export the relation as .txt.
3. Parse the file in Python and convert it into a structured dataframe.
4. Build the concept table.
5. Classify attributes according to reduction theory.
6. Compute consistent sets (reducts).
7. Validate implications between sets of attributes.

Implemented features

Concept analysis

• Construction of concept tables from object–attribute relations.
• Analysis of concept relationships.

Attribute classification

The project classifies attributes into:

• Infimum irreducibles (M_f)
• Infimum reducibles
• Absolutely necessary attributes (C_f)
• Relatively necessary attributes (K_f)
• Absolutely unnecessary attributes (I_f)

Reduct computation

Generation of:

• Consistent attribute sets
• Reducts

Implication validation

Implementation of:

implicacion_valida(atr1, atr2, df)

Checks whether one attribute set implies another over the given relation.

Technologies

• Python
• Pandas
• ConExp (for relation generation and concept lattices)

Example outputs

(Add screenshots here)

Input relation

Concept lattice

Program output

Dataframe

The concept table

The classification of attributes

Result of examples of implications

Repository structure

.
├── EXsept2024.txt (example)
├── images/
├── analisis_conceptos.ipynb
└── README.md

Academic context

This project combines mathematical theory and algorithmic implementation, focusing on formal concept analysis, attribute reduction and implication reasoning.