@MASTERSTHESIS\{IMM2012-06351,
    author       = "R. Troelsg{\aa}rd",
    title        = "Author and Topic Modelling in Text Data",
    year         = "2012",
    school       = "Technical University of Denmark, {DTU} Informatics, {E-}mail: reception@imm.dtu.dk",
    address      = "Asmussens Alle, Building 305, {DK-}2800 Kgs. Lyngby, Denmark",
    type         = "",
    note         = "Supervised by Professor Lars Kai Hansen, lkh@imm.dtu.dk, {DTU} Informatics",
    url          = "http://www.imm.dtu.dk/English.aspx",
    abstract     = "This thesis deals with probabilistic modelling of authors, documents, and topics in textual data. The focus is on the Latent Dirichlet Allocation (LDA) model and the Author-Topic (AT) model where Gibbs sampling is used for inferring model parameters from data. Furthermore, a method for optimising hyper parameters in an {ML-II} setting is described.
Model properties are discussed in connection with applications of the models which include detection of unlikely documents among scientific papers from the {NIPS} conferences using document perplexity, and the problem of link prediction in the online social network Twitter for which the results are reported as Area Under the {ROC} curve (AUC) and compared to well known graph-based methods."
}