Results 1  10
of
22
The maxmin hillclimbing bayesian network structure learning algorithm
 Machine Learning
, 2006
"... Abstract. We present a new algorithm for Bayesian network structure learning, called MaxMin HillClimbing (MMHC). The algorithm combines ideas from local learning, constraintbased, and searchandscore techniques in a principled and effective way. It first reconstructs the skeleton of a Bayesian n ..."
Abstract

Cited by 76 (7 self)
 Add to MetaCart
Abstract. We present a new algorithm for Bayesian network structure learning, called MaxMin HillClimbing (MMHC). The algorithm combines ideas from local learning, constraintbased, and searchandscore techniques in a principled and effective way. It first reconstructs the skeleton of a Bayesian network and then performs a Bayesianscoring greedy hillclimbing search to orient the edges. In our extensive empirical evaluation MMHC outperforms on average and in terms of various metrics several prototypical and stateoftheart algorithms, namely the PC, Sparse Candidate, Three Phase Dependency Analysis, Optimal Reinsertion, Greedy Equivalence Search, and Greedy Search. These are the first empirical results simultaneously comparing most of the major Bayesian network algorithms against each other. MMHC offers certain theoretical advantages, specifically over the Sparse Candidate algorithm, corroborated by our experiments. MMHC and detailed results of our study are publicly available at
Algorithms for Large Scale Markov Blanket Discovery
 In The 16th International FLAIRS Conference, St
, 2003
"... This paper presents a number of new algorithras for discovering the Markov Blanket of a target variable T from training data. The Markov Blanket can be used for variable selection for classification, for causal discovery, and for Bayesian Network learning. We introduce a loworder polynomial algorit ..."
Abstract

Cited by 28 (4 self)
 Add to MetaCart
This paper presents a number of new algorithras for discovering the Markov Blanket of a target variable T from training data. The Markov Blanket can be used for variable selection for classification, for causal discovery, and for Bayesian Network learning. We introduce a loworder polynomial algorithm and several variants that soundly induce the Markov Blanket under certain broad conditions in datasets with thousands of variables and compare them to other stateoftheart local and global methods with excellent results.
Searching for Bayesian Network Structures in the Space of Restricted Acyclic Aprtially Directed Graphs
 Journal of Artificial Intelligence Research
, 2003
"... Although many algorithms have been designed to construct Bayesian network structures using dierent approaches and principles, they all employ only two methods: those based on independence criteria, and those based on a scoring function and a search procedure (although some methods combine the two). ..."
Abstract

Cited by 15 (2 self)
 Add to MetaCart
Although many algorithms have been designed to construct Bayesian network structures using dierent approaches and principles, they all employ only two methods: those based on independence criteria, and those based on a scoring function and a search procedure (although some methods combine the two). Within the score+search paradigm, the dominant approach uses local search methods in the space of directed acyclic graphs (DAGs), where the usual choices for de ning the elementary modi cations (local changes) that can be applied are arc addition, arc deletion, and arc reversal. In this paper, we propose a new local search method that uses a dierent search space, and which takes account of the concept of equivalence between network structures: restricted acyclic partially directed graphs (RPDAGs). In this way, the number of dierent con gurations of the search space is reduced, thus improving eciency. Moreover, although the nal result must necessarily be a local optimum given the nature of the search method, the topology of the new search space, which avoids making early decisions about the directions of the arcs, may help to nd better local optima than those obtained by searching in the DAG space.
Bayesian Network Structure Learning by Recursive Autonomy Identification Raanan Yehezkel ∗ Video Analytics Group
"... We propose the recursive autonomy identification (RAI) algorithm for constraintbased (CB) Bayesian network structure learning. The RAI algorithm learns the structure by sequential application of conditional independence (CI) tests, edge direction and structure decomposition into autonomous substru ..."
Abstract

Cited by 3 (2 self)
 Add to MetaCart
We propose the recursive autonomy identification (RAI) algorithm for constraintbased (CB) Bayesian network structure learning. The RAI algorithm learns the structure by sequential application of conditional independence (CI) tests, edge direction and structure decomposition into autonomous substructures. The sequence of operations is performed recursively for each autonomous substructure while simultaneously increasing the order of the CI test. While other CB algorithms dseparate structures and then direct the resulted undirected graph, the RAI algorithm combines the two processes from the outset and along the procedure. By this means and due to structure decomposition, learning a structure using RAI requires a smaller number of CI tests of high orders. This reduces the complexity and runtime of the algorithm and increases the accuracy by diminishing the curseofdimensionality. When the RAI algorithm learned structures from databases representing synthetic problems, known networks and natural problems, it demonstrated superiority with respect to computational complexity, runtime, structural correctness and classification accuracy over the
Identifying Behavioral Principles underlying Activity Patterns by means of Bayesian Networks
, 2003
"... Capturing behavioral principles within the context of activitybased travel patterns is of vital importance to build adequate transportation planning models. Of course, there is no solitarily model which is perfectly capable of capturing all behavioral patterns but certain techniques are better suit ..."
Abstract

Cited by 2 (1 self)
 Add to MetaCart
Capturing behavioral principles within the context of activitybased travel patterns is of vital importance to build adequate transportation planning models. Of course, there is no solitarily model which is perfectly capable of capturing all behavioral patterns but certain techniques are better suited for it than others. In this paper the technique of Bayesian networks is introduced. Bayesian networks are potentially very strong representation techniques since they are capable of capturing the multidimensional nature of complex decisions. Several arguments are presented which clarify why the presented approach is particularly well suited to identify behavioral patterns. To this end and as an empirical study, several significant findings which might exert influence on the choice of transport mode choice were extracted from a large number of potential factors, in the context of a large activity diary dataset. Furthermore, the paper shows a detailed sensitivity analysis report which enables a quantitative evaluation.
AgentBased Distributed Intrusion Alert System,” to appear
 in Proc. 61h International Workshop on Distributed Computing
"... Abstract. Intrusion detection for computer systems is a key problem in today’s networked society. Current distributed intrusion detection systems (IDSs) are not fully distributed as most of them centrally analyze data collected from distributed nodes resulting in a single point of failure. Increasin ..."
Abstract

Cited by 2 (0 self)
 Add to MetaCart
Abstract. Intrusion detection for computer systems is a key problem in today’s networked society. Current distributed intrusion detection systems (IDSs) are not fully distributed as most of them centrally analyze data collected from distributed nodes resulting in a single point of failure. Increasingly, researchers are focusing on distributed IDSs to circumvent the problems of centralized approaches. A major concern of fully distributed IDSs is the high false positive rates of intrusion alarms which undermine the usability of such systems. We believe that effective distributed IDSs can be designed based on principles of coordinated multiagent systems. We propose an AgentBased Distributed Intrusion Alert System (ABDIAS) which is fully distributed and provides two capabilities in addition to other functionalities of an IDS: (a) early warning when preattack activities are detected, (b) detecting and isolating compromised nodes by trust mechanisms and votingbased peerlevel protocols. 1
Helping Countries Combat Corruption: Progress at the World Bank since
 IN: Electronic proceedings of the 83rd Annual Meeting of the Transportation Research Board
, 1997
"... * corresponding author ..."
Control in a 3D Reconstruction System using Selective Perception
"... This paper presents a control structure for general purpose image understanding that addresses both the high level of uncertainty in local hypotheses and the computational complexity of image interpretation. The control of vision algorithms is performed by an independent subsystem that uses Bayesian ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
This paper presents a control structure for general purpose image understanding that addresses both the high level of uncertainty in local hypotheses and the computational complexity of image interpretation. The control of vision algorithms is performed by an independent subsystem that uses Bayesian networks and utility theory to compute the marginal value of information provided by alternative operators and selects the ones with the highest value. We have implemented and tested this control structure with several aerial image datasets. The results show that the knowledge base used by the system can be acquired using standard learning techniques and that the valuedriven approach to the selection of vision algorithms leads to performance gains. Moreover, the modular system architecture simplifies the addition of both control knowledge and new vision algorithms.
unknown title
"... Inferring large gene networks from microarray data: a constraintbased approach We apply a constraintbased Bayesian network inference algorithm to the problem of discovering the network of genes involved in four types of lung carcinoma using microarray gene expression data. The large number of vari ..."
Abstract
 Add to MetaCart
Inferring large gene networks from microarray data: a constraintbased approach We apply a constraintbased Bayesian network inference algorithm to the problem of discovering the network of genes involved in four types of lung carcinoma using microarray gene expression data. The large number of variables (892), the small sample size (73 – typical for current microarray technology), as well as the noisy data require the ability to reconcile possibly unreliable conditional independence tests producing mutually inconsistent results. Our improved constraintbased algorithm QFCI is especially suited for inferring the global gene network structure (even in the presence of unknown hidden variables) rather than just fragmentary highscoring substructures. Moreover, QFCI was able to reconstruct a plausible substructure of the ‘small cell ’ subtype involving an expression profile typical for neuroendocrine differentiation. 1 Introduction and
Bayesian Decision Support in Medical Screening
"... It is expected that the availability of large data sets will lead to important changes in health care as these data can be exploited for the construction of decision support systems which may change the quality of patient care. Probabilistic graphical models, in particular Bayesian networks, are con ..."
Abstract
 Add to MetaCart
It is expected that the availability of large data sets will lead to important changes in health care as these data can be exploited for the construction of decision support systems which may change the quality of patient care. Probabilistic graphical models, in particular Bayesian networks, are considered appropriate tools for mining medical data. However, learning Bayesian networks (automatically) from data is time consuming and the quality of resulting structures is debatable. Research has shown that simple naive Bayesian classifiers often outperform sophisticated Bayesian networks for classification purposes. The latter have been shown to improve, though, by adding additional dependences between variables, if medical domain knowledge can be modelled properly. This project intends to develop new improved classifiers using Bayesian networks based on advanced image analysis and domain knowledge from the breast cancer screening domain to be used in decision support systems for radiologists. Recent breast cancer research has revealed that the reading of mammograms by radiologists is the weakest link in breast cancer screening. From audits it is known that in the Netherlands more than 25 % of all cancers detected in the screened population show relatively clear signs of abnormality in previous screening mammograms. Methods for computeraided detection (CAD) have been developed to support radiologists, but are usually intended to be used to avoid perception errors