Uncertainty in Bayesian leave-one-out cross-validation based model comparison
T Sivula, M Magnusson, AA Matamoros… - arXiv preprint arXiv …, 2020 - arxiv.org
… Here we have not used bias correction and assume that small bias doesn’t change the
general behaviour compared to LOO-CV. If K-fold-CV is used with grouped data division to …
general behaviour compared to LOO-CV. If K-fold-CV is used with grouped data division to …
Bayesian leave-one-out cross-validation approximations for Gaussian latent variable models
A Vehtari, T Mononen, V Tolvanen, T Sivula… - Journal of Machine …, 2016 - jmlr.org
The future predictive performance of a Bayesian model can be estimated using Bayesian
cross-validation. In this article, we consider Gaussian latent variable models where the …
cross-validation. In this article, we consider Gaussian latent variable models where the …
[PDF][PDF] Expectation propagation as a way of life: A framework for Bayesian inference on partitioned data
… Assume we have obtained at iteration t for node k, a set of posterior simulation draws θs t,k,
s = 1,...,… , possibly with weights ws t,k; take ws t,k ≡ 1 for an unweighted sample. To progress …
s = 1,...,… , possibly with weights ws t,k; take ws t,k ≡ 1 for an unweighted sample. To progress …
Expectation propagation as a way of life: A framework for Bayesian inference on partitioned data
… that is large enough, keep this sample, θs t,k+1 = θs t,k. Otherwise throw it away, simulate
new θs t+1,k’s from gt \k+1 , and reset the weights wt,k+1 to 1. This basic approach was used …
new θs t+1,k’s from gt \k+1 , and reset the weights wt,k+1 to 1. This basic approach was used …
Customer satisfaction measurement in emergency medical services
M Kuisma, T Määttä, T Hakala, T Sivula… - Academic …, 2003 - Wiley Online Library
Objectives: The annual patient volume in emergency medical services (EMS) systems is high
worldwide. However, there are no comprehensive studies on customer satisfaction for EMS…
worldwide. However, there are no comprehensive studies on customer satisfaction for EMS…
Machine Learning-Boosted Docking Enables the Efficient Structure-Based Virtual Screening of Giga-Scale Enumerated Chemical Libraries
T Sivula, L Yetukuri, T Kalliokoski… - Journal of Chemical …, 2023 - ACS Publications
The emergence of ultra-large screening libraries, filled to the brim with billions of readily
available compounds, poses a growing challenge for docking-based virtual screening. Machine …
available compounds, poses a growing challenge for docking-based virtual screening. Machine …
[HTML][HTML] Evolutionary aspects of inorganic pyrophosphatase
T Sivula, A Salminen, AN Parfenyev, P Pohjanjoki… - FEBS letters, 1999 - Elsevier
… found in all plant PPases close to the N-terminus, an insertion of one asparagine residue
next to the conserved glycine (G-82 in E-PPase) and an insertion in the region where T. …
next to the conserved glycine (G-82 in E-PPase) and an insertion in the region where T. …
Computational and experimental performance analysis of a novel method for heating of domestic hot water with a ground source heat pump system
T Niemelä, M Manner, A Laitinen, TM Sivula… - Energy and …, 2018 - Elsevier
… F t o t a l = Q h e a t , t o t a l Q e l e c t r i c i t y , H P + a u x . where: Q heat,total is the total
heating energy produced by the heat pump system during the measurement period, including …
heating energy produced by the heat pump system during the measurement period, including …
[HTML][HTML] Resolving outbreak dynamics using approximate Bayesian computation for stochastic birth–death models
J Lintusaari, P Blomstedt, B Rose, T Sivula… - Wellcome open …, 2019 - ncbi.nlm.nih.gov
… We give the net transmission rate t 1 a uniform prior over a large interval from 0 to 30.
Given the condition imposed by Inequality (3), we assign R 1 and R 2 uniform priors over a …
Given the condition imposed by Inequality (3), we assign R 1 and R 2 uniform priors over a …
Unbiased estimator for the variance of the leave-one-out cross-validation estimator for a Bayesian normal model with fixed variance
T Sivula, M Magnusson, A Vehtari - Communications in Statistics …, 2023 - Taylor & Francis
… analytic expectation of the naive estimator is calculated by applying equations (A53) and (A54)
from the appendix, in the following known expectation in a more general setting (Sivula, …
from the appendix, in the following known expectation in a more general setting (Sivula, …