Balanced sensitivity functions for tuning multi-dimensional Bayesian network classifiers

Bolt, Janneke H.; Gaag, Linda C. van der

doi:10.1016/j.ijar.2016.07.011

Cited by 18 publications

(21 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Because of powerful modeling capabilities of probabilistic graph model (PGM), most existing MDC approaches model class dependencies by assuming different DAG structures over class spaces. These models give rise to a family of PGMs for MDC called multi-dimensional Bayesian network classifiers [13,15,18,20]. Nonetheless, learning and inference in PGMs are computationally demanding.…”

Section: Related Workmentioning

confidence: 99%

“…Firstly, q 2 first-level classifiers are trained (steps 1-6). Then, training sets which will be used to train the second-level classifiers are constructed (steps [7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22]. After that, the second-level classifier for each class space is induced one by one (steps [23][24][25].…”

Section: The Seem Approachmentioning

confidence: 99%

“…It is widely acknowledged that modeling dependencies among class spaces is one of the core ways for designing better MDC approaches. Most existing approaches solve MDC problems by fitting probabilistic graphical model to MDC data, which can explicitly model dependencies among class spaces via assuming directed acyclic graph (DAG) structure over class spaces [13][14][15][16][17][18][19][20]. Dependencies also can be modeled by transforming MDC problem into a chain of multi-class classification problems [21,22], where chaining order over class spaces is critical but difficult to be determined.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Multi-dimensional classification via stacked dependency exploitation

Jia

Zhang

2020

Sci. China Inf. Sci.

View full text Add to dashboard Cite

Multi-dimensional classification (MDC) aims to build classification models for multiple heterogenous class spaces simultaneously, where each class space characterizes the semantics of an object w.r.t. one specific dimension. Modeling dependencies among class spaces plays a key role in solving MDC tasks, where most approaches work by assuming directed acyclic graph (DAG) structure or random chaining structure over class spaces. Different from existing probabilistic strategies, a deterministic strategy named Seem for dependency modeling is proposed in this paper via stacked dependency exploitation. In the first-level, pairwise dependencies are considered which can be modeled more reliably than modeling full dependencies among all class spaces by DAG or chaining structure. In the second-level, the class label of unseen instance w.r.t. each class space is determined by adaptively stacking predictive outputs from first-level pairwise classifiers. Experimental results show that stacked dependency exploitation leads to superior performance against stateof-the-art MDC approaches.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: The Seem Approachmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Multi-dimensional classification via stacked dependency exploitation

Jia

Zhang

2020

Sci. China Inf. Sci.

View full text Add to dashboard Cite

show abstract

“…Moreover, dependencies among class spaces can be explicitly modeled with directed acyclic graph (DAG) with different families of DAG structures (Bielza, Li, and Larrañaga 2011;Batal, Hong, and Hauskrecht 2013;Zhu, Liu, and Jiang 2016;Bolt and van der Gaag 2017;Benjumeda, Bielza, and Larrañaga 2018). Class powerset (CP) models dependencies by transforming the MDC problem into a single multi-class classification problem, where each possible combination of class variables y ∈ Y is treated as a new class in the transformed problem.…”

Section: Related Workmentioning

confidence: 99%

“…Nonetheless, dependencies among class spaces are ignored in this case which would impact the generalization performance of induced predictive model. Therefore, existing MDC approaches work by modeling dependencies among class variables from different dimensions in various ways, such as capturing pairwise interactions between class variables (Arias et al 2016), specifying chaining order over class variables (Zaragoza et al 2011;Read, Martino, and Luengo 2014), assuming directed acyclic graph (DAG) structure over class variables (Bielza, Li, and Larrañaga 2011;Batal, Hong, and Hauskrecht 2013;Zhu, Liu, and Jiang 2016;Bolt and van der Gaag 2017;Benjumeda, Bielza, and Larrañaga 2018), and partitioning class variables into groups (Read, Bielza, and Larrañaga 2014), etc.…”

Section: Introductionmentioning

confidence: 99%

Multi-Dimensional Classification via kNN Feature Augmentation

Jia¹,

Zhang²

2019

AAAI

View full text Add to dashboard Cite

Multi-dimensional classification (MDC) deals with the problem where one instance is associated with multiple class variables, each of which specifies its class membership w.r.t. one specific class space. Existing approaches learn from MDC examples by focusing on modeling dependencies among class variables, while the potential usefulness of manipulating feature space hasn’t been investigated. In this paper, a first attempt towards feature manipulation for MDC is proposed which enriches the original feature space with kNNaugmented features. Specifically, simple counting statistics on the class membership of neighboring MDC examples are used to generate augmented feature vector. In this way, discriminative information from class space is encoded into the feature space to help train the multi-dimensional classification model. To validate the effectiveness of the proposed feature augmentation techniques, extensive experiments over eleven benchmark data sets as well as four state-of-the-art MDC approaches are conducted. Experimental results clearly show that, compared to the original feature space, classification performance of existing MDC approaches can be significantly improved by incorporating kNN-augmented features.

show abstract

Structure-Based Categorisation of Bayesian Network Parameters

Bolt

Renooij

2017

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Bayesian networks typically require thousands of probability para-meters for their specification, many of which are bound to be inaccurate. Know-ledge of the direction of change in an output probability of a network occasioned by changes in one or more of its parameters, i.e. the qualitative effect of parameter changes, has been shown to be useful both for parameter tuning and in pre-processing for inference in credal networks. In this paper we identify classes of parameter for which the qualitative effect on a given output of interest can be identified based upon graphical considerations.

show abstract

Balanced sensitivity functions for tuning multi-dimensional Bayesian network classifiers

Cited by 18 publications

References 8 publications

Multi-dimensional classification via stacked dependency exploitation

Multi-dimensional classification via stacked dependency exploitation

Multi-Dimensional Classification via kNN Feature Augmentation

Structure-Based Categorisation of Bayesian Network Parameters

Contact Info

Product

Resources

About