🧠 Machine Learning & Deep Learning Projects

This repository contains two end-to-end projects demonstrating practical implementations of classical machine learning and deep learning techniques using real-world datasets.

---

🔬 1. Tumor Classification Using Traditional ML

Dataset: Breast Cancer Wisconsin Diagnostic Dataset (UCI)
Goal: Predict whether a tumor is benign or malignant based on diagnostic features.

🔹 Highlights:

• Built a full ML pipeline from data preprocessing to model deployment.
• Implemented and compared 7 different algorithms: SVM, KNN, Random Forest, Logistic Regression, Decision Tree, Naive Bayes, and Gradient Boosting.
• Evaluated models on accuracy, efficiency, and complexity.
• Achieved 96.1% accuracy, with detailed comparison across algorithms.
• Focused on model generalization, optimization, and interpretability.

Tools & Libraries:
Python, scikit-learn, pandas, matplotlib, seaborn, Google Colab

📄 Notebook: breast_cancer_model.ipynb

---

🧮 2. Handwritten Digit Recognition with CNN

Dataset: MNIST Handwritten Digit Dataset
Goal: Classify handwritten digits (0–9) using a Convolutional Neural Network.

🔹 Highlights:

• Designed a CNN architecture using TensorFlow and Keras for image classification.
• Achieved 98.7% validation accuracy, with strong generalization on unseen data.
• Applied preprocessing (normalization), used ReLU activations and softmax for output.
• Visualized training performance via loss/accuracy plots and feature map activations.

Tools & Libraries:
Python, TensorFlow, Keras, Matplotlib, NumPy

📄 Notebook: mnist_cnn_model.ipynb

---

📌 How to Use

1. Clone the repo:

   git clone https://github.com/shivigoyal4321/machine-learning.git

2. Launch notebooks in Jupyter or Google Colab. Install required libraries: pip install scikit-learn pandas matplotlib seaborn tensorflow keras

Shivi Goyal – Machine Learning Enthusiast