Machine Learning & Reinforcement Learning Algorithms

This repository contains a collection of Jupyter notebooks implementing fundamental machine learning, decision-making, and reinforcement learning algorithms. Each notebook focuses on a specific concept, combining theory with practical implementation for educational and experimental purposes.

---

📂 Repository Structure

code
├── Research Models
│ ├── multi-arm-bandit.ipynb
│ ├── q-learning.ipynb
│ ├── svm.ipynb
├── actor-critic.ipynb
├── bayesian-decision-making.ipynb
├── logistic-regression.ipynb
├── multi-arm-bandit.ipynb
├── neural-network.ipynb
├── policy-gradient.ipynb
├── q-learning.ipynb
└── svm.ipynb

---

Research Paper

The following algorithms are specifically used in the research paper:

1. Support Vector Machine (SVM)
2. Multi-Armed Bandit
3. Q-Learning

---

📘 Notebook Descriptions

1. Logistic Regression

   • Binary classification using logistic regression
   • Model formulation, training, and evaluation -Gradient-based optimization

2. Support Vector Machine (SVM)

   • Linear and margin-based classification
   • Decision boundaries and hinge loss
   • Practical implementation from scratch

3. Neural Network

   • Feedforward neural network implementation
   • Activation functions and backpropagation
   • Training and inference workflow

4. Bayesian Decision Making

   • Probabilistic reasoning under uncertainty
   • Bayesian inference and decision rules
   • Applications to optimal decision policies

5. Multi-Armed Bandit

   • Exploration vs. exploitation trade-off
   • ε-greedy and related strategies
   • Performance comparison of bandit algorithms

6. Q-Learning

   • Model-free reinforcement learning
   • Q-table updates and temporal-difference learning
   • Policy derivation from learned values

7. Policy Gradient

   • Direct policy optimization methods
   • Stochastic policies and gradient estimation
   • Reinforcement learning with function approximation

8. Actor-Critic

   • Hybrid value-based and policy-based approach
   • Actor and critic architecture
   • Advantage estimation and learning stability

---

🛠️ Requirements

To run the notebooks, ensure the following are installed:

• Python 3.8+
• Jupyter Notebook / JupyterLab
• NumPy
• Pandas
• Matplotlib
• (Optional) SciPy, scikit-learn

Install dependencies using:

    pip install numpy matplotlib scikit-learn jupyter

---

▶️ How to Run

1. Clone the repository:

       git clone https://github.com/TechMLW/QuantFP
       cd QuantFP

2. Launch Jupyter Notebook:

       jupyter notebook

3. Open any .ipynb file and run the cells sequentially.

---

🎯 Purpose

This repository is intended for:

• Learning core machine learning and reinforcement learning algorithms
• Academic coursework and self-study
• Experimentation with algorithmic concepts from scratch

---