Article
Computer Science, Interdisciplinary Applications
Polina Suter, Giusi Moffa, Jack Kuipers, Niko Beerenwinkel
Summary: BiDAG is an R package that implements Markov chain Monte Carlo (MCMC) methods for structure learning and sampling of Bayesian networks. It provides tools for finding the MAP graph and sampling graphs from the posterior distribution. A hybrid approach is used for inference in large graphs, with a reduced search space defined in the first step and an iterative order MCMC scheme optimizing the restricted search space in the second step. The package can handle both discrete and continuous data and includes MCMC schemes for structure learning and sampling of dynamic Bayesian networks.
JOURNAL OF STATISTICAL SOFTWARE
(2023)
Article
Statistics & Probability
Jack Kuipers, Polina Suter, Giusi Moffa
Summary: Bayesian networks are widely used probabilistic graphical models for understanding dependencies in high-dimensional data and facilitating causal discovery. This article proposes a novel hybrid method that combines constraint-based methods and score and search approaches to reduce the complexity of MCMC methods and achieve superior performance, enabling full Bayesian model averaging.
JOURNAL OF COMPUTATIONAL AND GRAPHICAL STATISTICS
(2022)
Article
Biochemical Research Methods
Polina Suter, Jack Kuipers, Niko Beerenwinkel
Summary: This study presents a strategy for learning gene regulatory networks (GRNs) using Dynamic Bayesian networks (DBNs) from gene expression data. The proposed approach is scalable, has high predictive accuracy, and prevents overfitting. The application of DBNs to two time series transcriptomic datasets demonstrates improved classification accuracy and the identification of differences in gene networks between cancer and normal tissues.
BRIEFINGS IN BIOINFORMATICS
(2022)
Article
Computer Science, Artificial Intelligence
George Bai, Rohitash Chandra
Summary: This paper presents a Bayesian ensemble learning framework that utilizes Bayesian inference to improve prediction accuracy and quantify uncertainty. By combining multiple base learners and training them using MCMC sampling, the framework has good scalability on large-scale models.
Article
Engineering, Multidisciplinary
Kenny Chowdhary, Chi Hoang, Kookjin Lee, Jaideep Ray, V. G. Weirs, Brian Carnes
Summary: This paper explores the effectiveness of combining machine-learning methods with projection-based model reduction techniques to create data-driven surrogate models of computationally expensive, high-fidelity physics models. The method is demonstrated on modeling heat flux and pressure in a turbulent flow and used for Bayesian estimation of parameters in a turbulence model for a high-fidelity flow simulator.
COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING
(2022)
Article
Statistics & Probability
Yves Atchade, Liwei Wang
Summary: This paper proposes a fast approximate Markov chain Monte Carlo sampling framework for a large class of sparse Bayesian inference problems. The computational cost per iteration in several regression models is of order O(n(s+J)), which can be further reduced by data sub-sampling. The algorithm is an extension of the asynchronous Gibbs sampler and can be viewed as a form of Bayesian iterated sure independent screening.
JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES B-STATISTICAL METHODOLOGY
(2023)
Article
Computer Science, Artificial Intelligence
Xianchang Wang, Hongjia Ren, Xiaoxin Guo
Summary: This paper presents a novel method that uses a discrete firefly optimization algorithm to learn the structure of a Bayesian network. Experimental results show that the proposed algorithm has better convergence accuracy and higher scores compared to other algorithms, indicating its effectiveness for learning Bayesian network structures.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Infectious Diseases
Tianyi Luo, Jiaojiao Wang, Quanyi Wang, Xiaoli Wang, Pengfei Zhao, Daniel Dajun Zeng, Qingpeng Zhang, Zhidong Cao
Summary: The study aims to reconstruct the complete transmission chain of the COVID-19 outbreak in Beijing's Xinfadi Market using epidemiological investigation data, contributing to understanding transmission dynamics and risk factors. Results show that the transmission rate of COVID-19 within households is 9.2%, and older people are more susceptible. The accuracy of the reconstructed transmission chain is 67.26%. In the Beef and Mutton Trading Hall of Xinfadi market, most transmission occurs within 20 meters, with an average transmission distance of 13.00 meters and the deepest transmission generation being the 9th.
INTERNATIONAL JOURNAL OF INFECTIOUS DISEASES
(2022)
Article
Physics, Multidisciplinary
Shouta Sugahara, Itsuki Aomi, Maomi Ueno
Summary: This study aims to improve classification accuracy using a modified subbagging method and demonstrates that it outperforms previous methods in the field of Bayesian network classification.
Article
Computer Science, Artificial Intelligence
Wei Fang, Weijian Zhang, Li Ma, Yunlin Wu, Kefei Yan, Hengyang Lu, Jun Sun, Xiaojun Wu, Bo Yuan
Summary: Bayesian networks (BNs) are powerful models for representation and reasoning under uncertainty. This paper presents a genetic algorithm-based approach called SIGA-BN for learning the structure of BNs. SIGA-BN utilizes the concepts of Markov blankets and v-structures to improve the learning process. The experimental results on benchmark networks demonstrate that SIGA-BN outperforms other GA-based and traditional BN structure learning algorithms in terms of structural accuracy, convergence speed, and computational time.
SWARM AND EVOLUTIONARY COMPUTATION
(2023)
Article
Computer Science, Artificial Intelligence
Martin Magris, Alexandros Iosifidis
Summary: The last decade has seen a growing interest in Bayesian learning, but its technicality and complexity in practical implementations have limited its widespread adoption. This survey introduces the principles and algorithms of Bayesian Learning for Neural Networks from a practical perspective, discussing standard and recent approaches for Bayesian inference. It also explores the use of manifold optimization as a state-of-the-art approach and provides pseudo-codes for implementation.
ARTIFICIAL INTELLIGENCE REVIEW
(2023)
Article
Computer Science, Information Systems
Rohitash Chandra, Ayush Bhagat, Manavendra Maharana, Pavel N. Krivitsky
Summary: Deep learning models, such as convolutional neural networks, have been widely used for image and multimedia tasks, with recent focus on unstructured data represented by graphs. Graph convolutional neural networks utilize graph-based data representation and convolutions for automatic feature extraction. Despite their popularity in various fields, uncertainty quantification remains a challenge.
Article
Computer Science, Artificial Intelligence
Baodan Sun, Yun Zhou, Jianjiang Wang, Weiming Zhang
Summary: This paper presents a heuristic algorithm combining PC and PSO algorithms for learning the structure of Bayesian networks, considering structure priors to enhance algorithm performance, and introducing new mutation and crossover operators. Experimental results demonstrate that the proposed approach outperforms other algorithms in terms of Bayesian Information Criterion (BIC) scores.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Baodan Sun, Yun Zhou
Summary: This paper proposes an improved biased random-key genetic algorithm to solve the BN structure learning problem. A local optimization model is applied as its decoder to enhance the algorithm's performance. Experimental results demonstrate that the proposed algorithm achieves better accuracy than other state-of-the-art algorithms and performs well in XSS attack detection for web security.
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
(2022)
Article
Mathematics
Francisco Javier Diez, Manuel Arias, Jorge Perez-Martin, Manuel Luque
Summary: OpenMarkov is an open-source software tool designed for probabilistic graphical models, primarily in medicine but also used in other fields and education in over 30 countries. This paper explains how OpenMarkov can be used as a pedagogical tool to teach the main concepts of Bayesian networks and influence diagrams, as well as various inference algorithms.
Article
Psychology, Clinical
Giusi Moffa, Jack Kuipers, Giuseppe Carra, Cristina Crocamo, Elizabeth Kuipers, Matthias Angermeyer, Traolach Brugha, Mondher Toumi, Paul Bebbington
Summary: Recent studies have found that affective symptoms may play a significant role in the onset of schizophrenic disorder, and persistent psychotic symptoms may be driven by the existence of affective disturbance. These findings have important implications for long-term treatment and interventions in psychiatry.
PSYCHOLOGICAL MEDICINE
(2023)
Article
Statistics & Probability
Jack Kuipers, Polina Suter, Giusi Moffa
Summary: Bayesian networks are widely used probabilistic graphical models for understanding dependencies in high-dimensional data and facilitating causal discovery. This article proposes a novel hybrid method that combines constraint-based methods and score and search approaches to reduce the complexity of MCMC methods and achieve superior performance, enabling full Bayesian model averaging.
JOURNAL OF COMPUTATIONAL AND GRAPHICAL STATISTICS
(2022)
Article
Surgery
Fabian Haak, Savas Soysal, Elisabeth Deutschmann, Giusi Moffa, Heiner C. Bucher, Max Kaech, Christoph Kettelhack, Otto Kollmar, Marco von Strauss Und Torney
Summary: The introduction of case load requirements (CR) for liver surgery in Switzerland in 2013 led to an increase in the number of liver resections, but had limited effects on centralization of care, according to a retrospective analysis of national in-hospital data.
WORLD JOURNAL OF SURGERY
(2022)
Article
Medicine, General & Internal
Andreas Michael Schmitt, Martin Walter, Amanda Katherina Herbrand, Markus Joerger, Giusi Moffa, Urban Novak, Lars Hemkens, Benjamin Kasenda
Summary: This study described the characteristics and survival of patients with cancer who had intended off-label use (OLU) cancer treatment and reimbursement request. The results showed that patients with access to intended OLU were younger, had better overall prognosis, less frequently had solid cancer, were in earlier stages, and had longer median overall survival compared to patients without access.
Article
Biochemical Research Methods
Polina Suter, Jack Kuipers, Niko Beerenwinkel
Summary: This study presents a strategy for learning gene regulatory networks (GRNs) using Dynamic Bayesian networks (DBNs) from gene expression data. The proposed approach is scalable, has high predictive accuracy, and prevents overfitting. The application of DBNs to two time series transcriptomic datasets demonstrates improved classification accuracy and the identification of differences in gene networks between cancer and normal tissues.
BRIEFINGS IN BIOINFORMATICS
(2022)
Article
Biochemical Research Methods
Anne Bertolini, Michael Prummer, Mustafa Anil Tuncel, Ulrike Menzel, Maria Lourdes Rosano-Gonzalez, Jack Kuipers, Daniel Johannes Stekhoven, Niko Beerenwinkel, Franziska Singer
Summary: Single-cell RNA sequencing (scRNA-seq) is a powerful technique for understanding tissue composition and disease mechanisms at the single-cell level. However, analyzing and interpreting the large amounts of data generated by scRNA-seq is challenging. In this study, we developed a workflow called scAmpi (Single Cell Analysis mRNA pipeline) that processes and analyzes scRNA-seq data from raw sequencing to provide clinically relevant information. The workflow removes low quality cells, identifies cell types, and visualizes gene expression and functional pathways in single cells. Additionally, scAmpi can link gene expression to potential drug candidates for disease treatment.
PLOS COMPUTATIONAL BIOLOGY
(2022)
Article
Computer Science, Interdisciplinary Applications
Polina Suter, Giusi Moffa, Jack Kuipers, Niko Beerenwinkel
Summary: BiDAG is an R package that implements Markov chain Monte Carlo (MCMC) methods for structure learning and sampling of Bayesian networks. It provides tools for finding the MAP graph and sampling graphs from the posterior distribution. A hybrid approach is used for inference in large graphs, with a reduced search space defined in the first step and an iterative order MCMC scheme optimizing the restricted search space in the second step. The package can handle both discrete and continuous data and includes MCMC schemes for structure learning and sampling of dynamic Bayesian networks.
JOURNAL OF STATISTICAL SOFTWARE
(2023)
Review
Health Care Sciences & Services
Pernilla Dillner, Luisa C. Eggenschwiler, Anne W. S. Rutjes, Lena Berg, Sarah N. Musy, Michael Simon, Giusi Moffa, Ulrika Forberg, Maria Unbeck
Summary: This study aimed to report the incidence and characteristics of adverse events (AEs) in paediatric inpatient care using different detection methods. The results showed a highly variable incidence of AEs, and the poor reporting standards and methodological differences hindered result comparison.
BMJ QUALITY & SAFETY
(2023)
Article
Medicine, General & Internal
Soheila Aghlmandi, Florian S. S. Halbeisen, Ramon Saccilotto, Pascal Godet, Andri Signorell, Simon Sigrist, Dominik Glinz, Giusi Moffa, Andreas Zeller, Andreas F. F. Widmer, Andreas Kronenberg, Julia Bielicki, Heiner C. C. Bucher
Summary: This randomized clinical trial found that quarterly personalized antibiotic prescribing audit and feedback with peer benchmarking did not reduce antibiotic prescribing among primary care physicians in Switzerland with medium to high antibiotic prescription rates.
JAMA INTERNAL MEDICINE
(2023)
Article
Biotechnology & Applied Microbiology
Senbai Kang, Nico Borgsmueller, Monica Valecha, Jack Kuipers, Joao M. Alves, Sonia Prado-Lopez, Debora Chantada, Niko Beerenwinkel, David Posada, Ewa Szczurek
Summary: This article presents a statistical method called SIEVE for the joint inference of somatic variants and cell phylogeny from single-cell DNA sequencing. The SIEVE method utilizes raw read counts for all nucleotides and corrects acquisition bias of branch lengths. It outperforms other methods in phylogenetic reconstruction and variant calling accuracy, especially for homozygous variants. When applied to three datasets, including triple-negative breast cancer and colorectal cancer samples, SIEVE finds that double mutant genotypes are rare in colorectal cancer but unexpectedly frequent in triple-negative breast cancer.
Article
Mathematics, Interdisciplinary Applications
Jack Kuipers, Giusi Moffa
Summary: Adjusting for covariates is a method to estimate causal effects, but different adjustment sets may lead to different precisions. The selection of adjustment set needs to consider the relationships between variables and sample size.
JOURNAL OF CAUSAL INFERENCE
(2022)