A collection of Machine Learning projects and algorithm implementations built to understand the mathematics, intuition, and engineering behind modern ML.
Rather than treating machine learning as a black box, this repository documents my journey of implementing algorithms from first principles, experimenting with datasets, and exploring how they work under the hood.
This repository serves as a central place for all my Machine Learning projects—from complete applications to algorithms implemented from scratch.
My focus is on:
- Understanding the mathematics behind ML
- Building intuition before using libraries
- Writing clean, well-documented implementations
- Experimenting with real-world datasets
- Tracking my learning journey publicly
ML-Projects-Scratch/
│
├── Cricket-Win-Probability/
│ ├── matches.csv
│ ├── probability_meter.py
│ └── README.md
│
├── Cancer-Detection-Classifier/
│ ├── Pipeline.py
│ └── README.md
│
└── README.md
Each project contains its own documentation, implementation details, dataset information, and results.
| Project | Description | Status |
|---|---|---|
| 🏏 Cricket Win Probability | Predicts the probability of winning an IPL match using machine learning. | ✅ Completed |
| Cancer Detection Classifier | PCA and SVM implemented from scratch using covariance matrices and Hinge Loss. | ✅ Completed |
- Python
- NumPy
- Pandas
- Matplotlib
- Scikit-learn (only for datasets/evaluation where appropriate)
I believe that understanding why an algorithm works is more valuable than simply knowing how to use it.
Whenever possible, I implement algorithms from scratch before relying on high-level libraries.
This repository is an ongoing record of that journey.
Some of the upcoming implementations include:
- DBSCAN
- HDBSCAN
- Naive Bayes
- Gradient Boosting
- XGBoost
- Neural Networks
- CNNs
- Transformers
- Reinforcement Learning
I'm always open to suggestions, discussions, and improvements.
If you notice something that can be implemented more efficiently or explained better, feel free to open an issue or start a discussion.
If you found this repository useful or interesting, consider giving it a ⭐.
I'm always happy to connect with fellow students, developers, and machine learning enthusiasts!