Statistical Methods for ICT Applications


            Arnab Bhattacharya

Arnab Bhattacharya

Contact Details: School of Computer Science and Statistics,
Lloyd Institute (Room 105),
Trinity College,
Dublin 2,

+353 (0) 1 896 2048
+353 (0) 1 677 0711


    Summary of Research:

    The problem of performing online inference when observations arrive sequentially in time is well known in the statistical and engineering communities.  Statisticians have come up with several state-of-the-art techniques to make inferences to online problems, while keeping a balance between accuracy and speed.  My current research focuses on developing a new technique that would be able to cater for high dimensionality and speed at the same time.

    I am trying to develop a functional approximation to prediction of general state space models, where the latent variable can be high dimensional and parameters can be assumed to be unknown.  I plan to extend this is a spatial setting in the future.

    General interests:  I have a strong interest in Bayesian Statistics and its applications.  I am also particulary interested in classical Time Series and regression, both linear and non-linear.  State-Space modelling is my current area of interest.                            

    Publications and Presentations:

    B. Houlding, A. Battacharya, S. P. Wilson, and T. K. Forde, 'A fast Bayesian model for latent radio signal prediction'.  Technical Report (09/02) , Trinity College Dublin, 2009.