KTH Mathematics  


Timo Koski

  • Papers 2008-- :

    • J. Corander, Y. Cui, & T. Koski: Inductive Inference and Partiton Exchangeability in Classifiction. Ray Solomonoff Conference, Lecture Notes in Computer Science (in press)

    • J. Corander, Y. Cui, T. Koski & J. Siren: Have I Seen You Before ? Principles of Bayesian Predictive Classification Revisited. Statistics and Computing Link

    • M. Ohlson & Koski, T. (2011). On the Distribution of Matrix Quadratic Forms. Accepted for publication in Communications in Statistics - Theory and Methods (CIS).
    • J. Corander, M. Gyllenberg & T.Koski(2010): Learning Genetic Population Structures Using Minimization of Stochastic Complexity. Entropy; Link

    • J. Corander, M.Ekdahl & T.Koski (2009): Bayesian Unsupervised Learning of DNA Regulatory Binding Regions (in Advances in Artificial Intelligence 2009) Link

    • J. Corander, M.Gyllenberg & T.Koski (2009): Bayesian unsupervised classification framework based on stochastic partitions of data and a parallel search strategy (in Advances in Data Analysis and Classification 2009) Link

    • J. Corander, M.Ekdahl & T.Koski (2008): Parallell interacting MCMC for learning of graph topologies (in Data Mining and Knowledge Discovery 2008) Link


  • A preprint on measurement of research productivity (2011):


  • Papers (2003-2007) via LiU Research Database:

    • The complete list (two pages) Link


  • Computer science library:


  • Recent selected papers from LiU Research Database:

    • M.Ekdahl & T.Koski (2007): On Concentration of Discrete Distributions with Applications to Supervised Learning of Classifiers Link

    • J. Corander, M. Gyllenberg, & T.Koski (2007): Random Partition Models and Exchangeability for Bayesian Identification of Population StructureLink

    • M.Ekdahl & T.Koski (2006): On Concentration of Discrete Distributions with Applications to Supervised Learning of ClassifiersBounds for the Loss in Probability of Correct Classification Under Model Based Approximation Link


  • Preprints/technical reports:

    • Timo Koski and Roland Orre (1998): Statistics of the Information Component in Bayesian Neural Networks. pdf

    • Mats Gyllenberg and Timo Koski (2001): Probabilistic Models for Bacterial Taxonomy Link

    • Mats Gyllenberg, Timo Koski and Tatu Lund (2001): BinClass: A Software Package for Classifying Binary Vectors User's Guide Link

    • Linus Göransson and Timo Koski (2002): Using a Dynamic Bayesian Network to Learn Genetic Interactions. pdf

    • Mats Gyllenberg, Jonas Carlsson and Timo Koski (2002): Bayesian Network Classification of Binarized DNA Fingerprint Patterns. ps

    • Harry Hurd and Timo Koski (2002): The Wold isomorphism for cyclostationary sequences. pdf



  • A news item from 'Microbiology` :




    Senast uppdaterad 2007-11-23.