Algorithms+for+Time+Series+Analysis

Algorithms for Time Series Analysis

**Organizer**: Pavlos Protopapas

 This workshop will focus on algorithms for Time Series Analysis and cover four major subjects in two 90 minute sessions. Each talk/tutorial including questions and discussion w ill be for 45 minutes.

**A. Classification. **//Speaker:// Ashish Mahabal (Caltech). Automatic classification using machine learning methods is becoming a standard in surveys with large datasets. This session will review various methods and present few examples of such applications. **B. Time Series Modeling. **//Speaker//: Suzanne Aigrain (Oxford). This session will cover autoregressive models and multivariate approaches such as Gaussian Processes. **C. Meta-classification/mixture of expert models. **//Speaker//: Karim Pichara (Pontificia Universidad Católica from Chile). Given that machine learning classification methods are showing great promise in automatic classification, this session will cover how we combine various methods together. **D. Event Detection ****.** //Speaker//: Pavlos Protopapas (Harvard). This session will address methods of fast identification of  low signal-noise-ratio events, characterization and statistical issus of low-signal-noise ratio and rare events.