A New Ensemble Learning Algorithm Combined with Causal Analysis for Bayesian Network Structural Learning

Li, Ming; Zhang, Ren; Liu, Kefeng

doi:10.3390/sym12122054

Cited by 12 publications

(4 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…After the DBN network structure is established, the network parameters can be learned based on the discrete indicator samples and the network structure, that is, the influence relationships among network nodes are quantitatively mined and expressed in the form of conditional probability. Parameter learning requires the determination of observation conditional probability and transition conditional probability (Li et al 2020). We adopt the learning algorithm based on DS evidence theory, supplemented by variation coefficient (VC) method, Monte Carlo (MC) method and EM algorithm, to learn the conditional probability distribution of nodes.…”

Section: Parameter Learningmentioning

confidence: 99%

Assessment of underwater navigation safety based on dynamic Bayesian network facing uncertain knowledge and various information

Zhang²,

Chen³

et al. 2022

Front. Mar. Sci.

Self Cite

View full text Add to dashboard Cite

As ocean environment is complicated and varied, underwater vehicles (UVs) are facing great challenges in safe and precise navigation. Therefore, it is important to evaluate the underwater ocean environment safety for the UV navigation. To deal with the uncertain knowledge and various information in the safety assessment, we present an evaluation model based on the dynamic Bayesian network (DBN) theory. Firstly, characteristic indicators are extract from marine environment systems and discretized with Cloud model. Then, the DBN is constructed through structure learning and parameter learning based on Dempster-Shafer (DS) evidence theory. Finally, the dynamic evaluation and risk zoning of the navigation safety is realized based on Bayesian probabilistic reasoning. The DBN-based assessment model fully considers the uncertainty of influence relationships between marine environment and UV navigation, and effectively fuses expert knowledge and quantitative data for assessment modeling. The experimental results show the proposed model has high reliability and good value of application.

show abstract

Section: Parameter Learningmentioning

confidence: 99%

Assessment of underwater navigation safety based on dynamic Bayesian network facing uncertain knowledge and various information

Zhang²,

Chen³

et al. 2022

Front. Mar. Sci.

Self Cite

View full text Add to dashboard Cite

show abstract

“…In the domain of causal reasoning specifically, it is known that more reasonable causal relationships can be extracted through the combination process of ensemble learning [16], which addresses the inconsistency of different causal inference algorithms when applied to the same input. Such causality ensembles have previously been investigated by Li et al [17], who utilized a bagging mechanism with a new weighting criterion to fuse different Bayesian network (BN) structures. This framework has higher accuracy and a more powerful generalization ability than a single BN-learner.…”

Section: Introductionmentioning

confidence: 99%

A Data-Driven Two-Phase Multi-Split Causal Ensemble Model for Time Series

Kemmerling

Buschmann

et al. 2023

Symmetry

View full text Add to dashboard Cite

Causal inference is a fundamental research topic for discovering the cause–effect relationships in many disciplines. Inferring causality means identifying asymmetric relations between two variables. In real-world systems, e.g., finance, healthcare, and industrial processes, time series data from sensors and other data sources offer an especially good basis to infer causal relationships. Therefore, many different time series causal inference algorithms have been proposed in recent years. However, not all algorithms are equally well-suited for a given dataset. For instance, some approaches may only be able to identify linear relationships, while others are applicable for non-linearities. Algorithms further vary in their sensitivity to noise and their ability to infer causal information from coupled vs. non-coupled time series. As a consequence, different algorithms often generate different causal relationships for the same input. In order to achieve a more robust causal inference result, this publication proposes a novel data-driven two-phase multi-split causal ensemble model to combine the strengths of different causality base algorithms. In comparison to existing approaches, the proposed ensemble method reduces the influence of noise through a data partitioning scheme in a first phase. To achieve this, the data are initially divided into several partitions and the base causal inference algorithms are applied to each partition. Subsequently, Gaussian mixture models are used to identify the causal relationships derived from the different partitions that are likely to be valid. In the second phase, the identified relationships from each base algorithm are then merged based on three combination rules. The proposed ensemble approach is evaluated using multiple metrics, among them a newly developed evaluation index for causal ensemble approaches. We perform experiments using three synthetic datasets with different volumes and complexity, which have been specifically designed to test causality detection methods under different circumstances while knowing the ground truth causal relationships. In these experiments, our causality ensemble outperforms each of its base algorithms. In practical applications, the use of the proposed method could hence lead to more robust and reliable causality results.

show abstract

“…Through systematic analysis of risk-causing factors and riskbearing bodies of marine disasters, Li et al (2018b) adopted the BN and gray relational analysis to build a risk assessment model. Then, Li et al (2020;2021a) also proposed the improved weighted BN to mine the causal relationship of disaster factors and realized probabilistic reasoning of marine disaster risk.…”

Section: Introductionmentioning

confidence: 99%

“…Therefore, BN learning needs to be supported by large-scale samples. However, natural disasters are extreme events with very small occurring frequency and there are few observation samples, which makes it difficult to learn the BN structure and parameters from data sets (Li et al, 2020). In existing studies about BNbased risk assessment, the network structure and parameters are manually constructed by experts based on domain knowledge.…”

Section: Introductionmentioning

confidence: 99%

Expert Knowledge-Driven Bayesian Network Modeling for Marine Disaster Assessment Under the Small Sample Condition

Zhang

Liu

2022

Front. Mar. Sci.

Self Cite

View full text Add to dashboard Cite

Risk assessment and management of marine disasters are the prerequisite of ocean exploitation and utilization. Marine disaster assessment is a complicated system engineering with high non-linearity and uncertainty. To deal with the problem, Bayesian network (BN) has become a powerful model used for disaster assessment due to its capability of expressing complex relationships and reasoning with uncertainty. However, scarce data sets and case samples of marine disasters pose an obstacle to BN modeling, particularly for structure and parameter learning. In our research, we combine expert knowledge with small sample to propose a new BN-based assessment model. Expert knowledge is regularly expressed and quantitatively incorporated into BN learning with DS evidence theory. Then, the genetic algorithm is adopted to search the optimal network parameters. Comparative experiments show that the new model has a better assessment accuracy (91.03%) than BPNN (61.34%) and SVM (70.67%), especially with small samples. The proposed model achieves the risk assessment of marine disasters under the small sample condition, providing the technical support for marine disaster prevention and mitigation.

show abstract

A New Ensemble Learning Algorithm Combined with Causal Analysis for Bayesian Network Structural Learning

Cited by 12 publications

References 20 publications

Assessment of underwater navigation safety based on dynamic Bayesian network facing uncertain knowledge and various information

Assessment of underwater navigation safety based on dynamic Bayesian network facing uncertain knowledge and various information

A Data-Driven Two-Phase Multi-Split Causal Ensemble Model for Time Series

Expert Knowledge-Driven Bayesian Network Modeling for Marine Disaster Assessment Under the Small Sample Condition

Contact Info

Product

Resources

About