Rule-Based Learning Engine

This repository contains a dataset-driven, explainable concept learning system implemented using the Find-S algorithm (Machine Learning – Unit 1).

The project focuses on learning decision rules from structured data and applying them to real-time API-style JSON inputs for security-related decisions.

---

Image

---

Problem Statement

Traditional machine learning projects often rely on black-box models that are difficult to interpret.
This project demonstrates how concept learning can be used to build transparent, explainable decision systems using structured datasets.

The system learns a hypothesis from data and uses it to make decisions on unseen inputs received via APIs.

---

Learning Algorithm

• Find-S Algorithm
  • Learns the most specific hypothesis consistent with all positive examples
  • Maintains explainability through explicit hypothesis representation
  • Reflects inductive bias toward consistency over probability

No probabilistic models or external ML libraries are used.

---

Dataset Design

• Custom-designed categorical dataset (30+ records)
• Features inspired by real-world email security indicators:
  • URL presence
  • Sender reputation
  • Keyword score
  • Capital usage ratio
  • Attachment type
• Labels :
  • Spam
  • Legitimate

The dataset is intentionally categorical to align with concept learning theory.

---

🔄 System Architecture