Time Series Classification at Scale

This thesis develops scalable algorithms and techniques to classify large amount of time series data. Nowadays, many real-world applications are generating huge amount of time series data. This wealth of data is required to create finer and more accurate classification models that allow us to learn from the data. Unfortunately, the state-of-the-art classification algorithms are impractical for large amount of time series data. Models that are accurate but slow are not good. Therefore, the ability to classify large amount of data quickly and accurately allows the state of the art to be more practical in many applications.