QCG Open Project 2025: Quantum Machine Learning Classifier

📋 Project Overview

This project is a part of the QCG Open Projects 2025. It demonstrates the implementation of a Variational Quantum Classifier (VQC) to solve binary classification problems. By leveraging quantum feature maps and entangling layers, the model learns to classify complex data by mapping it into a high-dimensional Hilbert space.

The repository includes implementations using standard healthcare data (Diabetes dataset) to showcase the practical utility of Hybrid Quantum-Classical Algorithms.

---

📂 Repository Structure

File	Description
qml.ipynb	The core notebook containing the VQC architecture, circuit training, and initial results.
qml_newdataset.ipynb	Extension notebook testing the model robustness on a secondary dataset.
diabetes.csv	Primary dataset containing diagnostic features for classification.
dataset.csv	Supplemental dataset used for benchmarking and validation.

---

🧠 Technical Workflow

1. Data Pre-processing

Classical data from diabetes.csv is normalized using Min-Max Scaling to ensure feature values are compatible with quantum gate rotation ranges ($[0, \pi]$ or $[0, 2\pi]$).

2. Quantum Feature Mapping

We utilize Angle Encoding or ZZFeatureMap to transform classical vectors into quantum states $|\psi\rangle$. This step is crucial for capturing non-linear relationships that are difficult for classical kernels to identify.

3. Variational Circuit (Ansatz)

The trainable part of the circuit consists of:

• Rotation Layers: Parameterized $R_y$ and $R_z$ gates.
• Entanglement Layers: CNOT gates to create quantum correlations between features.

4. Optimization

A classical optimizer (e.g., COBYLA or Adam) (COBYLA) in this case is used to iteratively update the circuit parameters by minimizing the Cross-Entropy loss calculated from quantum measurements.

---

🚀 Getting Started

Prerequisites

• Python 3.9+
• Jupyter Notebook or VS Code

Installation

1. Clone this repository:

   git clone [https://github.com/Rasp05-ops/qcg_project.git](https://github.com/Rasp05-ops/qcg_project.git)
   cd qcg_project