0000000000070010
AUTHOR
Santeri Karppinen
Valkosolupitoisuuksien bayesilainen mallintaminen lasten leukemian ylläpitohoidossa
Lasten akuutin lymfoblastileukemian ylläpitovaiheen hoidossa tehtävät lääkeannostuspäätökset pohjataan nykyisin potilaan veren valkosolupitoisuuteen, joka on hoidon tehokkuudesta kertova tekijä. Potilaalle sopiva lääkeannostus on hoidon onnistumisen ja turvallisuuden kannalta tärkeä, mutta sen löytäminen on vaikeaa, sillä annettu lääkitys näkyy valkosolupitoisuudessa viiveellä, ja potilaiden elimistön reagointi lääkitykseen on yksilöllistä. Sopivan lääkeannostuksen löytämistä hankaloittavat myös hoidonaikaiset tulehdukset, jotka voivat muuttaa valkosolupitoisuutta hetkellisesti. Työ käsittelee akuuttiin lymfoblastileukemiaan sairastuneiden suomalaisten potilaiden veren valkosolupitoisuuden …
Identifying territories using presence-only citizen science data : An application to the Finnish wolf population
Citizens, community groups and local institutions participate in voluntary biological monitoring of population status and trends by providing species data e.g. for regulations and conservation. Sophisticated statistical methods are required to unlock the potential of such data in the assessment of wildlife populations. We develop a statistical modelling framework for identifying territories based on presence-only citizen science data. The framework can be used to jointly estimate the number of active animal territories and their locations in time. Our approach is based on a data generating model which consists of a dynamic submodel for the appearance/removal of territories and an observatio…
Conditional particle filters with bridge backward sampling
Conditional particle filters (CPFs) with backward/ancestor sampling are powerful methods for sampling from the posterior distribution of the latent states of a dynamic model such as a hidden Markov model. However, the performance of these methods deteriorates with models involving weakly informative observations and/or slowly mixing dynamics. Both of these complications arise when sampling finely time-discretised continuous-time path integral models, but can occur with hidden Markov models too. Multinomial resampling, which is commonly employed with CPFs, resamples excessively for weakly informative observations and thereby introduces extra variance. Furthermore, slowly mixing dynamics rend…
Conditional particle filters with diffuse initial distributions
Conditional particle filters (CPFs) are powerful smoothing algorithms for general nonlinear/non-Gaussian hidden Markov models. However, CPFs can be inefficient or difficult to apply with diffuse initial distributions, which are common in statistical applications. We propose a simple but generally applicable auxiliary variable method, which can be used together with the CPF in order to perform efficient inference with diffuse initial distributions. The method only requires simulatable Markov transitions that are reversible with respect to the initial distribution, which can be improper. We focus in particular on random-walk type transitions which are reversible with respect to a uniform init…
Prediction of leukocyte counts during paediatric acute lymphoblastic leukaemia maintenance therapy
Maintenance chemotherapy with oral 6-mercaptopurine and methotrexate remains a cornerstone of modern therapy for acute lymphoblastic leukaemia. The dosage and intensity of therapy are based on surrogate markers such as peripheral blood leukocyte and neutrophil counts. Dosage based leukocyte count predictions could provide support for dosage decisions clinicians face trying to find and maintain an appropriate dosage for the individual patient. We present two Bayesian nonlinear state space models for predicting patient leukocyte counts during the maintenance therapy. The models simplify some aspects of previously proposed models but allow for some extra flexibility. Our second model is an ext…