GISC6323 - AI for Socio-Economic Data

GISC 6323 (EPPS 6326) AI for Socio-Economic Data (3 semester credit hours) Key AI methods of supervised and unsupervised learning are covered. First core principles, such as feature transformation, the variance-bias tradeoff, sampling strategies, dimensionality reduction, and the Bayesian theorem, are discussed using associated machine learning methods. Second, support vector machines and ensemble tree-based methods are presented. Within the framework of neural networks and deep learning numerical predictions, basic text processing, convolutional image operations, autoencoders, and basic gradient optimization are introduced. Lastly, several unsupervised cluster algorithms, including spatial constraints and the interpretation of their groupings, are discussed. Using open-source software, the concepts and methods are illustrated using social sciences and geo-referenced data. (3-0) P