Unsupervised Machine Learning
This course introduces unsupervised machine learning as a practical framework for finding structure in complex data without predefined labels. Students work with clustering, dimensionality reduction, embeddings, matrix factorization, topic modeling, anomaly detection, graph representations, and model validation.
The course is designed for data science and computational social science applications: behavioral traces, text, networks, educational data, cultural data, organizations, and public-policy problems where the goal is to discover patterns, compare latent groups, and build interpretable representations.
The public course site is available at nosupervisado.criss-lab.com.
Cristian Candia
Associate Professor, Data Science Institute, School of Engineering, Universidad del Desarrollo, Chile. Head of CRiSS-LAB.
Cristian Candia studies how societies transform information into collective relevance through attention, memory, preferences, and coordination. His work combines computational social science, network science, AI, and large-scale behavioral data to understand how groups, institutions, and societies decide what matters.