Dr. Patrick Schäfer - Unleashing the Potential of Unsupervised Time Series Analytics: Recent Advances and Breakthroughs
Low-cost, high-resolution sensors, as equipped in smart-phones and watches, have become increasingly
popular, generating vast amounts of raw time series data. This includes accelerometer, sound, or
temperature data. However, supervised machine learning techniques for analyzing time series require
labeled data, which is often not available. Fortunately, unsupervised analytics offer valuable tools to
extract insights and enhance our understanding of the data. In this lecture, we will discuss recent
advances in unsupervised time series analysis that uncover hidden patterns and gain new insights without
relying on labels.