Direct Multi-label Linear Discriminant Analysis

Οικονόμου, Μαρία; Tefas, Anastasios

doi:10.1007/978-3-642-41013-0_43

Cited by 8 publications

(3 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For comparisons, we used the following LDA-based dimensionality reduction techniques: DMLDA [21], wMLDAc, wMLDAb, wMLDAe, wMLDAf, and wMLDAd [14], where the subscripts denote the types of prior information used as weight factors following Section III-A1. Note that wMLDAc is equivalent to the original MLDA [12].…”

Section: Methodsmentioning

confidence: 99%

“…In [21], MLDA was extended to Direct MLDA by changing the definition of S b in a way that allows obtaining a higher dimensional subspace than the original MLDA, where the subspace dimensionality is limited by the rank of S b to C−1. This extension work further enhanced the results in multilabel video classification tasks.…”

Section: B Linear Discrimination Analysis-based Algorithms For Multimentioning

confidence: 99%

“…Correlation-based prior information (SMLDAc) was used as weight factors in the original MLDA algorithm [12]. As in [12], we first compute the label correlation matrix R defined in (21). We then compute the normalized weight vector m j ∈ R C using ( 22) for all data items and set our prior information matrix values as…”

Section: A Probabilistic Multilabel Class-saliency Estimationmentioning

confidence: 99%

See 2 more Smart Citations

Saliency-Based Multilabel Linear Discriminant Analysis

Raitoharju

Iosifidis

et al. 2022

IEEE Trans. Cybern.

View full text Add to dashboard Cite

Linear discriminant analysis (LDA) is a classical statistical machine-learning method, which aims to find a linear data transformation increasing class discrimination in an optimal discriminant subspace. Traditional LDA sets assumptions related to the Gaussian class distributions and single-label data annotations. In this article, we propose a new variant of LDA to be used in multilabel classification tasks for dimensionality reduction on original data to enhance the subsequent performance of any multilabel classifier. A probabilistic class saliency estimation approach is introduced for computing saliency-based weights for all instances. We use the weights to redefine the between-class and within-class scatter matrices needed for calculating the projection matrix. We formulate six different variants of the proposed saliency-based multilabel LDA (SMLDA) based on different prior information on the importance of each instance for their class(es) extracted from labels and features. Our experiments show that the proposed SMLDA leads to performance improvements in various multilabel classification problems compared to several competing dimensionality reduction methods.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: B Linear Discrimination Analysis-based Algorithms For Multimentioning

confidence: 99%