Introduction to Machine Learning on Data Streams Using River

Learn how to leverage River, a Python library for machine learning on streaming data, in this 32-minute conference talk from PyCon South Africa. Explore techniques for handling continuous, unbounded data streams generated from diverse sources like network devices, cloud computing, wireless systems, financial transactions, and environmental monitoring. Discover how to implement machine learning tasks including regression, classification, and unsupervised learning using River as an extension of the Massive Online Analysis (MOA) framework. Master essential concepts like online metrics computation, concept drift detection, and building effective machine learning pipelines for evaluating model performance on streaming datasets. Gain practical insights into processing real-time data from various domains including healthcare monitoring, electricity usage prediction, and surveillance systems.