Learning Oncogenetic Networks by Reducing to Mixed Integer Linear Programming

At present, most of the algorithms for learning Bayesian Networks (BNs) use EM algorithm to deal with incomplete data. They are of low efficiency because EM algorithm has to perform iterative process of probability reasoning to complete the incomplete data. In this paper we present an efficient BN learning algorithm, which use the combination of EMI method and a scoring function based on mutual information theory. The algorithm first uses EMI method to estimate, from incomplete data, probability distributions over local structures of BNs, then evaluates BN structures with the scoring function and searches for the best one. The detailed procedure of the algorithm is depicted in the paper. The experimental results on Asia and Alarm networks show that when achieving high accuracy, the algorithm is much more efficient than two EM based algorithms, SEM and EM-EA algorithms.

113 GENOMIC PROFILING OF RENAL CELL CARCINOMA IN PATIENTS WITH END-STAGE RENAL DISEASE

Bayesian networks can be used to model gene regulatory networks because of its capability of capturing causal relationships between genes. However, learning Bayesian network is an NP-hard problem. Hill climbing methods are used in BN learning, in which K2 is a frequently used greedy search algorithm. But the performance of K2 algorithm is greatly affected by a prior ordering of input nodes and relatively low accuracy of the learned structures may be observed. To solve these problems, we propose a new algorithm (BPSOBN) to explore the use of Binary Particle Swarm Optimization (BPSO) algorithms for learning Bayesian networks. The result of experiments show that our BPSO based algorithm can obtain better networks than hill climbing methods. BPSOBN also shows the effectiveness for network reconstruction to gene expression data measured during the yeast cell cycle.

Bayesian networks have become a popular knowledge representation scheme for probabilistic knowledge. One of the main challenges in Bayesian networks structure learning is the development of inductive learning techniques that scale up to large and possibly physically distributed data sets. We consider an approach to learning the structure of BN from distributed data. This is based on the collective learning strategy, where a local structure is obtained at each site and the global structure is obtained by cross learning and combining. Local learning is used to identify the local structure of local nodes and local links of cross nodes from local sample data sets. Cross learning can find the entire cross-links of cross nodes. Combined together, the local BNs and BN learnt from cross learning and removes any extra local links. We use the collective learning algorithm in distributed flood decision supporting system. The result shows that the Bayesian network structure is the same compare the collective learning algorithm from distributed data sets with learning from concentrated data sets. And meanwhile the collective learning algorithm is more efficient than transmitting all samples to a central site.

Existing Structural Expectation-Maximization (EM) algorithms for learning Bayesian networks from incomplete data usually adopt the greedy hill climbing search method, which may make the algorithms find sub-optimal solutions. In this paper, we present a new Structural EM algorithm which employs a hybrid evolutionary algorithm as the search method. The experimental results on the data sets generated from several benchmark networks illustrate that our algorithm outperforms some state-of-the-art learning algorithms.

We present an improved method for learning Bayesian network (BN) structures. The new approach incorporate the idea of simulated annealing algorithm (SA) into the selection operator of genetic algorithm(GA). The BN structure with the highest score is given high priority in selection, as well as the structure with lower score could also be given opportunity to be selected. That is high score prior genetic-simulated annealing algorithm to Bayesian network structure learning(GSA_BNSL).This strategy will reserve optimal gene while avoiding the premature caused by the misleading from high score individual in the population. Experiments on comparison of several BN learning algorithms are carried out using typical data set of data mining. The result indicates that the GSA_BNSL is able to obtain an optimized BN structure with higher accuracy rate.

User's personal social networks are big and cluttered, yet contain highly valuable information. Organizing users' friends into circles or communities is a fundamental task in social network research. Social network sites allow users to manually categorize their friends into social circles, however this process is laborious and inadaptable to changes. In this paper, we study novel ways of automatically determining users' social circles. We treat this task as a classification problem on a user's ego-network, a network of connections between friends. Based on Bayesian Network (BN), we develop a model for determining whether a query user Uq is in main user Um's social circle. First, we transform the original social network data to make it suitable for BN modeling, and build an Initial Bayesian Network (IBN) of Um using the state-of-the-art BN learning algorithm. Then, we propose a new method to improve the IBN by adding important parents to the class variable. Lastly, leveraging carefully designed threshold, we use the final BN to determine the existence of Uq in the social circle of U <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">m</sub> . Modeling social circle with BN allows us to quantify user's social circle existence with probability and run query with missing values/evidences. Using ground-truth data from Facebook and Twitter, experimental results indicate that our BN model could accurately determine user's existence in social circle and outperforms four baseline predictors, namely Naïve Bayes, IBL, OneR and J48, showing promising application potential in the social circle research area.

Chapter 11 - Molecular Phylogenetics

The transcriptional response to exogenously supplied nitric oxide in Saccharomyces cerevisiae was modeled using an integrated framework of Bayesian network learning and experimental feedback. A Bayesian network learning algorithm was used to generate network models of transcriptional output, followed by model verification and revision through experimentation. Using this framework, we generated a network model of the yeast transcriptional response to nitric oxide and a panel of other environmental signals. We discovered two environmental triggers, the diauxic shift and glucose repression, that affected the observed transcriptional profile. The computational method predicted the transcriptional control of yeast flavohemoglobin YHB1 by glucose repression, which was subsequently experimentally verified. A freely available software application, ExpressionNet, was developed to derive Bayesian network models from a combination of gene expression profile clusters, genetic information and experimental conditions.

A new 3D object retrieval approach is proposed based on a novel Bayesian networks lightfield descriptor (BLD). To overcome the disadvantages of the existing 3D object retrieval methods, firstly, we explore Bayesian network for building a new lightfield descriptor, 3D object is put into lightfield, and multi-views information can be obtained along a sphere, and then features of images can be extracted out. The BLD of 3D object would be got according to feature sequences based on Bayesian network learning algorithm. Secondly, 3D object is retrieved based on graph model measurement and relevant feedback learning. Beneficial from the statistical learning and graph model optimization, proposed BLD is robustness as compared to the existing methods. Experimental results demonstrate that proposed approach is with better performance than the existing methods.

Various algorithms have been developed for variant calling using next-generation sequencing data, and various methods have been applied to reduce the associated false positive and false negative rates. Few variant calling programs, however, utilize the pedigree information when the family-based sequencing data are available. Here, we present a program, FamSeq, which reduces both false positive and false negative rates by incorporating the pedigree information from the Mendelian genetic model into variant calling. To accommodate variations in data complexity, FamSeq consists of four distinct implementations of the Mendelian genetic model: the Bayesian network algorithm, a graphics processing unit version of the Bayesian network algorithm, the Elston-Stewart algorithm and the Markov chain Monte Carlo algorithm. To make the software efficient and applicable to large families, we parallelized the Bayesian network algorithm that copes with pedigrees with inbreeding loops without losing calculation precision on an NVIDIA graphics processing unit. In order to compare the difference in the four methods, we applied FamSeq to pedigree sequencing data with family sizes that varied from 7 to 12. When there is no inbreeding loop in the pedigree, the Elston-Stewart algorithm gives analytical results in a short time. If there are inbreeding loops in the pedigree, we recommend the Bayesian network method, which provides exact answers. To improve the computing speed of the Bayesian network method, we parallelized the computation on a graphics processing unit. This allowed the Bayesian network method to process the whole genome sequencing data of a family of 12 individuals within two days, which was a 10-fold time reduction compared to the time required for this computation on a central processing unit.

Due to the uncertain nature of many of the factors that influence on the performance of an emergency medical service, we propose using Bayesian networks to model this kind of systems. We use an algorithm for learning Bayesian networks to build the model, from the point of view of a hospital manager, and apply it to the specific case of a spanish hospital. We also report the results of some preliminary experimentation with the model.KeywordsBayesian networksLearning algorithmsEmergency medical servicesManagement problems

Clinical reasoning in veterinary medicine is often based on clinicians' personal experience in combination with information derived from publications describing cohorts of patients. Studies on the use of scientific methods for patient individual decision making are largely lacking. This applies to the prediction of the individual underlying pathology in seizuring dogs as well. The aim of this study was to apply machine learning to the prediction of the risk of structural epilepsy in dogs with seizures. Dogs with a history of seizures were retrospectively as well as prospectively included. Data about clinical history, neurological examination, diagnostic tests performed as well as the final diagnosis were collected. For data analysis, the Bayesian Network and Random Forest algorithms were used. A total of 33 features for Random Forest and 17 for Bayesian Network were available for analysis. The following four feature selection methods were applied to select features for further analysis: Permutation Importance, Forward Selection, Random Selection and Expert Opinion. The two algorithms Bayesian Network and Random Forest were trained to predict structural epilepsy using the selected features. A total of 328 dogs of 119 different breeds were identified retrospectively between January 2017 and June 2021, of which 33.2% were diagnosed with structural epilepsy. An overall of 89,848 models were trained. The Bayesian Network in combination with the Random feature selection performed best. It was able to predict structural epilepsy with an accuracy of 0.969 (sensitivity: 0.857, specificity: 1.000) among all dogs with seizures using the following features: age at first seizure, cluster seizures, seizure in last 24 h, seizure in last 6 month, and seizure in last year. Machine learning algorithms such as Bayesian Networks and Random Forests identify dogs with structural epilepsy with a high sensitivity and specificity. This information could provide some guidance to clinicians and pet owners in their clinical decision-making process.

Uncertainty Modeling of a Group Tourism Recommendation System Based on Pearson Similarity Criteria, Bayesian Network and Self-Organizing Map Clustering Algorithm

Exact algorithms for learning optimal Bayesian networks require much more time and are used in small Bayesian networks. This paper adds ancestral partition constraints into the bidirectional heuristic search algorithm based on the order graph. The ancestral partition can be obtained by extracting strongly connected components from possible parent sets. Experiments show that ancestral partition can significantly improve the efficiency and scalability of bidirectional heuristic search. In addition, with ancestral partition constraints, bidirectional heuristic search has better efficiency and can search larger Bayesian networks than state-of-the-art algorithms.