Feb 28th, 2014, 3:00 pm in Huxley 245

Crouching Time Series, Hidden Markov Model - Applications of HMMs in the Real World

I will present an introduction and brief discussion into the applications and variations of the Hidden Markov Model (HMM), combining unsupervised learning techniques with performance analysis measures. Their parsimonious nature and efficient training on discrete and continuous data traces have made them popular as storage workloads, Markov Arrival Processes (MAPs), social network behaviour classifiers and financial predictive models (to name but a few). We explain relevant findings of the AESOP group (aesop.doc.ic.ac.uk/) over the last few years and mention possible future research.

Tiberiu Chis is a PhD student at Imperial College London in the AESOP (Analysis, Engineering, Simulation & Optimization of Performance) group under the supervision of Prof. Peter Harrison More about me→