This repository contains lab materials for the IST-AAut (Machine Learning) course.
- Check Report.pdf for the methodology and analysis, with comparisons of models across different metrics.
This section covers regression analysis tasks using synthetic data to explore key concepts in predictive modeling.
-
ML-Submission1: Contains the Jupyter Notebook for Multiple Linear Regression with Outliers, implementing outlier removal, cross-validation, and tuning techniques as described in the report.
-
ML-Submission2: Contains the Jupyter Notebook for the ARX Model, focusing on time-series data and system response modeling, with parameter optimization techniques.
- Understand multiple linear regression with synthetic data containing noise and outliers.
- Apply ARX (Auto-Regressive with eXogenous input) models for time-series data analysis.
- Evaluate model robustness with cross-validation and tuning techniques.
- Python (3.11) with
scikit-learnandstatsmodels. - MATLAB® for additional analysis and model validation.
This section explores image classification and segmentation tasks focused on low-resolution (48x48) Martian crater analysis.
-
ML-Submission3: Contains the Jupyter Notebook for Image Classification using SVC and CNN models, with techniques for handling imbalanced data and data augmentation.
-
ML-Submission4: Contains the Jupyter Notebook for Image Segmentation, implementing MLP-Fusion and U-Net models for pixel-wise segmentation.
- Develop machine learning models to classify crater vs. non-crater images.
- Apply segmentation techniques (patch-based and pixel-based) to delineate crater boundaries.
- Address data imbalance with techniques like SMOTE and data augmentation.
- Python (3.11) with
torch,torchvision,torchmetrics, andpytorch-lightning. - Optuna for hyperparameter tuning.
This work is licensed under a Creative Commons Attribution Non Commercial Share Alike 4.0 International.