SoleXplorer - A symbolic journey through your datasets

SoleXplorer.jl is an interactive interface for exploring symbolic machine learning models, built on top of the Sole.jl ecosystem. It provides tools for visualizing, inspecting, and interacting with models derived from (logic-based) symbolic learning algorithms.

---

Installation

using Pkg
Pkg.add("https://github.com/aclai-lab/SoleXplorer.jl/")

---

Quick Start

Decision tree

using SoleXplorer

# Load example dataset
dataset = load_sole_dataset("iris")

# Train a decision tree (uses `DecisionTree.jl`)
model = fit_tree(dataset, :species)

# Explore the model
explorer = Explorer(model, dataset)
run(explorer)

Temporal association rules

using SoleXplorer

# Load a temporal dataset
dataset = load_sole_dataset("temporal_events")

# Train a modal association ruleset (uses `ModalAssociationRules.jl`)
model = fit_modal_association_rules(dataset, :event_label; support=0.1, confidence=0.8)

# Start exploration
explorer = Explorer(model, logiset)
run(explorer)

In the exploration sessions, you can:

• View rules and their metrics
• Filter by modal depth or conditions
• Inspect logical formulas and their evaluation

---

Under the Hood: How It Works

1. Feature extraction Data instances are seen as models of logical formalisms (see SoleLogics.jl), and are represented in an optimized form for model checking (i.e., as logisets, see SoleData.jl).

2. Model Fitting Symbolic models (e.g. decision trees, modal association rules) are trained via SoleModels.jl-compliant packages.

3. Post-hoc Analysis Metrics and human-readable logic representations are obtained via rule extraction algorithms from SolePostHoc.jl.

4. Exploration Interface SoleXplorer.jl ties everything into an interactive exploration session, allowing you to inspect rules, formulas, and patterns from your symbolic datasets and models.

---

About

The package is developed by the ACLAI Lab @ University of Ferrara.