Machine Learning Assignments

This repository contains implementations of various Machine Learning algorithms and deep learning models applied to different datasets. The projects cover classification, regression, and unsupervised learning tasks.

Project Structure

1. Autoencoder

An unsupervised learning project using Neural Networks for dimensionality reduction (image embedding) and image restoration on the MNIST dataset.

2. KNN Classifier

Implementation of the K-Nearest Neighbors algorithm applied to two different domains:

• MNIST Dataset: Handwritten digit classification with performance analysis across different k values.
• Wine Dataset: Classification of wine varieties including data normalization and PCA visualization.

3. Naive Bayes Classifier

Implementation of the Categorical Naive Bayes algorithm for prediction tasks:

• Breast Cancer Prediction: Predicting cancer recurrence events.
• Weather Prediction: Predicting the decision to play sports based on weather conditions.

4. Neural Network Regression

A regression task estimating the health status (Compressor and Turbine decay) of Naval Propulsion Plants. Includes workflows for:

• Data preprocessing and normalization.
• Model training with multi-start optimization.
• Hyperparameter tuning using Cross-Validation.

Requirements

• Python 3.x
• TensorFlow / Keras
• Scikit-learn
• Pandas
• NumPy
• Matplotlib