Time series classification in resource constrained environments

Abstract

The analysis of time series is becoming prevalent across various scientific and engineering disciplines, where the e↵ectiveness and scalability of time series mining techniques depend on the design choices made while representing, indexing and comparing the time series. A lot of the existing algorithms in the field of time series classification can be resource intensive, making it difficult to apply them in an IoT environment, where we have several constraints on resources, with the need to process data being streamed from multiple sensors at the same time. The primary aim of this project is to implement an online time series classification algorithm which can work even in constrained environments.