Combination of Supervised and Unsupervised Classification Using the Theory of Belief Functions

Karem, Fatma; Dhibi, Mounir; Martin, Arnaud

doi:10.1007/978-3-642-29461-7_10

Cited by 13 publications

(8 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This paper proposes a new approach which improves a previous work [9]. The goal is to construct a fusion mechanism robust to noise,lack of sampling data and missing data.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Credal Fusion of Classifications for Noisy and Uncertain Data

Karem

Dhibi

Atkinson

et al. 2017

IJECE

Self Cite

View full text Add to dashboard Cite

This paper reports on an investigation in classification technique employed to classify noised and uncertain data. However, classification is not an easy task. It is a significant challenge to discover knowledge from uncertain data. In fact, we can find many problems. More time we don't have a good or a big learning database for supervised classification. Also, when training data contains noise or missing values, classification accuracy will be affected dramatically. So to extract groups from data is not easy to do. They are overlapped and not very separated from each other. Another problem which can be cited here is the uncertainty due to measuring devices. Consequentially classification model is not so robust and strong to classify new objects. In this work, we present a novel classification algorithm to cover these problems. We materialize our main idea by using belief function theory to do combination between classification and clustering. This theory treats very well imprecision and uncertainty linked to classification. Experimental results show that our approach has ability to significantly improve the quality of classification of generic database.

show abstract

“…This paper proposes a new approach which improves a previous work [9]. The goal is to construct a fusion mechanism robust to noise,lack of sampling data and missing data.…”

Section: Discussionmentioning

confidence: 99%

“…This work is an improvement of a previous one. The former [9] was established to combine clustering and classification in order to improve their performance. Both has difficulties.…”

Section: Fusion Mechanismmentioning

confidence: 99%

Credal Fusion of Classifications for Noisy and Uncertain Data

Karem

Dhibi

Atkinson

et al. 2017

IJECE

Self Cite

View full text Add to dashboard Cite

show abstract

“…The transformation of the mass function in pignistic probability is made in order to make a decision for the value of . 71 (5) where represents the cardinal of . Determine a credal partition from data returns to find for each individual , the quantities such that is greater if the distance between and focal element is small.…”

Section: Theory Of Belief Functionsmentioning

confidence: 99%

“…The belief functions introduced in the Dempster Shafer theory [4] have been widely used to model the uncertain and imprecise information for data classification and information fusion as example [5][6].…”

Section: Introductionmentioning

confidence: 99%

Kernel Credal C-Means Method Based on Belief Functions

Bendadi¹,

Lakhdar²,

Sbai

et al. 2017

IJMUE

View full text Add to dashboard Cite

The new approaches based on the credal partition commit objects; not only to singleton clusters but also to meta-clusters; with different masses of beliefs. Among these approaches we find the Credal C-means (CCM). In this article, we introduce a new clustering algorithm called kernel Credal C-means (KCCM), which is a kernel version of CCM algorithm. It is based on the beliefs functions. A second contribution of this paper is an improved CCM algorithm using a non-Euclidean metric unlike CCM which is based on Euclidean distance. In order to show the effectiveness of the proposed method, artificial and real data (vehicle trajectory in a bend) are tested at the end of this paper. A comparison is made to evaluate the results obtained with some methods as Evidential Cmeans (ECM) and CCM.

show abstract

“…The combination in the decision level of independent supervised and unsupervised classifiers is still a big challenge. In [15], the authors applied Dempster-Shafer theory to achieve this combination, considering supervised learning and unsupervised learning as independent sources, through which the uncertainty and the imprecision in the fusion can be measured. However, currently existing methods do not focus on the conflict situation where the number of supervised classes is not the same as the number of unsupervised clusters.…”

Section: Introductionmentioning

confidence: 99%

Combination of Supervised Learning and Unsupervised Learning Based on Object Association for Land Cover Classification

Martin

Estival

2018

2018 Digital Image Computing: Techniques and Applications (DICTA)

Self Cite

View full text Add to dashboard Cite

Conventional supervised classification approaches have significant limitations in the land cover classification from remote sensing data because a large amount of high quality labeled samples are difficult to guarantee. To overcome this limitation, combination with unsupervised approach is considered as one promising candidate. In this paper, we propose a novel framework to achieve the combination through object association based on Dempster-Shafer theory. Inspired by object association, the framework can label the unsupervised clusters according to the supervised classes even though they have different numbers. The proposed framework has been tested on the different combinations of commonly used supervised and unsupervised methods. Compared with the supervise methods, our proposed framework can furthest enhance the overall accuracy approximately by 8.2%. The experiment results proved that our proposed framework has achieved twofold performance gain: better performance on the insufficient training data case and the possibility to apply on a large area. Index Terms-Object association, land cover classification, combination of supervised learning and unsupervised learning.

show abstract

Combination of Supervised and Unsupervised Classification Using the Theory of Belief Functions

Cited by 13 publications

References 8 publications

Credal Fusion of Classifications for Noisy and Uncertain Data

Credal Fusion of Classifications for Noisy and Uncertain Data

Kernel Credal C-Means Method Based on Belief Functions

Combination of Supervised Learning and Unsupervised Learning Based on Object Association for Land Cover Classification

Contact Info

Product

Resources

About