Course

MICRO-455: Machine learning I

Lectures in this course (60)

Delves into advanced data preprocessing techniques, covering categorical encoding, missing data handling, and unbalanced datasets, emphasizing performance metrics and classifier comparison.

Machine Learning Fundamentals

Covers key concepts and examples of machine learning algorithms and techniques.

Pitfalls and Caveats in Machine Learning

Covers challenges in machine learning, emphasizing the importance of choosing relevant data and algorithms.

Kernel K-Means: Convergence Proof

Explores the Kernel K-Means algorithm, convergence proof, RBF kernel influence, and clustering interpretation.

Advanced Machine Learning: Fundamentals and Applications

Covers the fundamentals of advanced machine learning, emphasizing practical applications through interactive exercises and projects.

Pitfalls and Caveats in Machine Learning

Covers the importance of data and algorithms in machine learning and how to avoid noise.

Principal Component Analysis: Dimensionality Reduction

Explores Principal Component Analysis for dimensionality reduction in machine learning, showcasing its feature extraction and data preprocessing capabilities.

PCA: Intuition

Covers the basics of PCA, exercises on dimensionality reduction, and criteria for choosing projections.

PCA: Derivation and Optimization

Covers the derivation of PCA projection, error minimization, and eigenvector optimization.

Clustering Principle: Feature Extraction and Similarity Measures

Covers clustering for feature extraction, similarity measures, outliers, and cluster shapes.

K-means Clustering: Assignment and Update Steps

Explains the assignment and update steps in K-means clustering, loss function minimization, and distance metric effects.

Soft K-means Clustering & DBSCAN

Covers Soft K-means Clustering and DBSCAN principles, algorithms, and comparison.

Evaluation for Clustering

Covers the evaluation of clustering methods, including K-means clustering and the use of evaluation metrics to determine the optimal number of clusters.

Probability Distributions: Discrete and Continuous

Covers discrete and continuous probability distributions, including joint and conditional probabilities.

Fitting data with one Gauss function

Explains Gaussian functions, modeling data, likelihood function, and maximum likelihood optimization.

Fitting and Clustering Data with Mixture of Gauss Functions

Covers Mixture of Gauss Functions, Gaussian Mixture Modeling, and hyper-parameter optimization.

Probabilistic Interpretation: K-Means & GMM

Explores the probabilistic interpretation of K-means clustering and its relation to Gaussian Mixture Models.

Classification: Introduction

Covers clustering, semi-supervised clustering, and binary classification formalization, along with various classification techniques.

Classification with GMM

Explores the use of Gaussian Mixture Models for transitioning from clustering to classification, covering binary classification, parameter estimation, and optimal Bayes classifier.

KNN Classifier: Nearest Neighbor Approach

Explains the K-Nearest Neighbors classifier, assigning labels based on closest points and smoothing noise in labels.