MasterThesis_Projects

This repository contains the Jupyter Notebooks used for my Master Thesis Project at the University of Lisbon with the title: "Identifying subtypes of Mild Cognitive Impairment due to Alzheimer’s disease in cerebrospinal fluid proteomics: a Machine Learning approach"

Overview

The work focused on applying supervised and unsupervised Machine Learning methods to omics data for the identification of Alzheimer’s disease subtypes. The implemented pipelines include steps for:

Feature_Ranking_Ensemble.ipynb

• Data preprocessing and cleaning

• Feature ranking and selection using multiple strategies:

  • Univariate Selection
  • Single Score Methods
  • Attribute Weights from ML models
  • Decrease of Accuracy
  • Recursive Feature Addition (RFA)
  • Recursive Feature Elimination (RFE)

• Ensemble feature ranking, aggregating multiple selection results to build a more robust and stable feature set

• Supervised modeling and evaluating classification algorithms such as:

  • Support Vector Machines (SVM)
  • Naïve Bayes`
  • k-Nearest Neighbor (kNN)
  • Decision Trees
  • Logistic Regression
  • Random Forests
  • XGBoost

Feature_Clustering.ipynb

• Unsupervised clustering of patients using Non-negative Matrix Factorization (NMF) to identify molecular subtypes.

• Visualization and interpretation of model outputs to extract biologically meaningful insights

Publication:

The findings from this project are published in:

• Pathophysiological subtypes of mild cognitive impairment due to Alzheimer's disease identified by CSF proteomics