Time Series Prediction by Andreas S. Weigend

Forecasting the Future and Understanding the Past

An interdisciplinary collection that advances forecasting and interpretation of temporal data through nonlinear dynamics and machine learning, with a strong focus on neural networks. Anchored by a benchmark prediction competition, it contrasts classical statistics with embedding-based, local, Bayesian, and information-theoretic approaches, spotlighting overfitting, noise, chaos, and out-of-sample validation. It distills practical guidelines and empirical lessons for building robust predictors and extracting structure from complex real-world signals.

Purchase from Bookshop.org