Feature Selection for Multi-label Learning: A Systematic Literature Review and Some Experimental Evaluations

Spolaôr, Newton; Lee, Huei; Takaki, Weber Shoity Resende; Wu, Feng

doi:10.1080/18756891.2015.1129587

Cited by 8 publications

(9 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Several FS methods have been analyzed in order to select features from each sample, including Information Gain, Fscore, Relief, mutual information, and normalized mutual information. Based on the existing literature review [27] [30], these methods and their extensions prove to be effective in case of multi-label FS, and in addition, they are able to cope with the feature-label correlation [1].…”

Section: A Baseline Feature Selection Methodsmentioning

confidence: 99%

“…Relief is the most effective and commonly used FS in multi-label text classification system [27] [30].…”

Section: ) Mutual Information (Mi)mentioning

confidence: 99%

“…R, as a feature evaluation measure, more often selects smaller number of features than the other FS methods, without degrading the performance of the classifiers. This could be due to the fact that R considers interactions among the features [27].…”

Section: B Experiments Ii: Evaluation Of the Proposed Ensemble Fs Andmentioning

confidence: 99%

See 2 more Smart Citations

A Dynamic Two-Layers MI and Clustering-based Ensemble Feature Selection for Multi-Labels Text Classification

Taha¹,

Tiun²,

Rahman³

et al. 2020

IJACSA

View full text Add to dashboard Cite

Multi-label text classification deals with the issue that arises from each sample being related to multiple labels. The text data suffers from high dimensionality. In order to resolve this issue, a feature selection (FS) method can be implemented for efficiently removing the noisy, irrelevant, and redundant features. Multi-label FS is a powerful tool for solving the highdimension problem. With regards to handling correlation and high dimensionality problems in multi-label text classification, this paper investigates the various heterogeneous FS ensemble schemes. In addition, this paper proposes an enhanced FS method called dynamic multi-label two-layers MI and clusteringbased ensemble feature selection algorithm (DMMC-EFS). The proposed method considers the: 1) dynamic global weight of feature, 2) heterogeneous ensemble, and 3) maximum dependency and relevancy and minimum redundancy of features. This method aims to overcome the high dimensionality of multilabel datasets and acquire improved multi-label text classification. We have conducted experiments based on three benchmark datasets: Reuters-21578, Bibtex, and Enron. The experimental results show that DMMC-EFS has significantly outperformed other state-of-the-art conventional and ensemble multi-label FS methods.

show abstract

Section: A Baseline Feature Selection Methodsmentioning

confidence: 99%

“…Relief is the most effective and commonly used FS in multi-label text classification system [27] [30].…”

Section: ) Mutual Information (Mi)mentioning

confidence: 99%

See 1 more Smart Citation

A Dynamic Two-Layers MI and Clustering-based Ensemble Feature Selection for Multi-Labels Text Classification

Taha¹,

Tiun²,

Rahman³

et al. 2020

IJACSA

View full text Add to dashboard Cite

show abstract

“…Notwithstanding, Spolaôr et al in [35] and [36] have proposed extensions of ReliefF towards the task of MLC by redefining the distance function. More precisely, they propose to use normalized Hamming loss or Jaccard dissimilarity as distance functions between two examples.…”

Section: Related Workmentioning

confidence: 99%

HMC-ReliefF: Feature ranking for hierarchical multi-label classification

et al. 2018

View full text Add to dashboard Cite

In machine learning, the growing complexity of the available data poses an increased challenge for its analysis. The rising complexity is both in terms of the data becoming more high-dimensional as well as the data having a more intricate structure. This emphasizes the need for developing machine learning algorithms that are able to tackle both the high-dimensionality and the complex structure of the data. Our work in this paper focuses on the development and analysis of the HMC-ReliefF algorithm, which is a feature relevance (ranking) algorithm for the task of Hierarchical Multi-label Classification (HMC). The basis of the algorithm is the RReliefF algorithm for regression that is adapted for hierarchical multi-label target variables. We perform an extensive experimental investigation of the HMC-ReliefF algorithm on several datasets from the domains of image annotation and functional genomics. We analyse the algorithm's performance in terms of accuracy in a filterlike setting and also in terms of ranking stability for various parameter values. The results show that the HMC-ReliefF can successfully detect relevant features from the data that can be further used for constructing accurate predictive models. Additionally, the stability analysis helps to determine the preferred parameter values for obtaining not just accurate, but also a stable algorithm output.

show abstract

“…In the early days of relational databases, performance issues were extensive because of limited hardware resources and immature optimizers, so performance was a priority consideration. But even today, despite the huge growth in resources, there is even more growth in the amount of data available due to technological advances [1] and the amount and variety of connected devices [2], so performance continues to be of critical importance.…”

Section: Introductionmentioning

confidence: 99%

Techniques for Ensuring Index Usage Predictability in Microsoft SQL Server

et al. 2021

View full text Add to dashboard Cite

The demand for carrying out high-performance operations with data is growing in parallel with the vast growth of data itself. The retrieval of data for analysis, the manipulation of data, as well as its insertion in data stores must all be performed very efficiently, using techniques that ensure speed, reliability and accuracy. This paper investigates the techniques and practices that improve the performance of data retrieving by the use of SQL and Microsoft SQL Server. Being that SQL is a declarative language that specifies what should be produced as a result, instead of how to achieve that result, this paper will look at the internals of SQL Server that affect the "how" of queries and data operations, in order to propose techniques that ensure performance gains. The paper will aim to shed light on the limitations and variance in index usage, and to answer the question why indexes are sometimes used, and other times not, for the same query. To overcome the index limitations the "index fusion" technique is proposed.

show abstract

Feature Selection for Multi-label Learning: A Systematic Literature Review and Some Experimental Evaluations

Cited by 8 publications

References 17 publications

A Dynamic Two-Layers MI and Clustering-based Ensemble Feature Selection for Multi-Labels Text Classification

A Dynamic Two-Layers MI and Clustering-based Ensemble Feature Selection for Multi-Labels Text Classification

HMC-ReliefF: Feature ranking for hierarchical multi-label classification

Techniques for Ensuring Index Usage Predictability in Microsoft SQL Server

Contact Info

Product

Resources

About