Unsupervised learning techniques will be introduced steadily. Clustering can improve asset grouping in portfolio construction or signal classification; Anomaly detection can complement existing risk monitoring systems. and dimensionality reduction methods similar to PCA can improve model interpretability or data preprocessing. Crucially, they will complement slightly than replace existing models, making integration more viable and fewer disruptive. For investment practitioners, these methods enable tasks similar to regime discovery, portfolio diversification, signal classification and anomaly detection by revealing complex relationships and latent aspects which are often invisible to traditional approaches.
This chapter begins by introducing clustering methods, including -means, spectral clustering, and hierarchical clustering, and highlights their use in grouping assets, identifying market regimes, and constructing diversified portfolios. Notable use cases include: De Prado’s hierarchical risk parity framework and applications of spectral clustering to macroregime classification. The chapter then discusses dimensionality reduction techniques similar to PCA (t-SNE) and ICA as methods to simplify high-dimensional data sets.
