"One vs All" Classifier Analysis for Multi-label Movie Genre Classification Using Document Embedding.

Sonia Guehria; Habiba Belleili; Nabiha Azizi; Samir Brahim Belhaouari

doi:10.1007/978-3-030-71187-0_44

"One vs All" Classifier Analysis for Multi-label Movie Genre Classification Using Document Embedding.

Sonia Guehria, Habiba Belleili, Nabiha Azizi, Samir Brahim Belhaouari

CSE Information & Computing Technology

Badji Mokhtar University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

The classification of movie genres from their synopses has attracted
the attention of many researchers. Indeed, synopses are a source of relevant
information that contributes to determinate movie genre. The automation of this
classification process is very useful in several applications, such recommenda-
tion systems. Moreover, movies can belong simultaneously to several genres
(drama, action, comedy, horror), which reflects a typical problem of multi-label
classification (MLC). In this article, we use a powerful representation of film
synthesis via a document integration technique Doc2vec in the multi-label
classification context. The technique used in our experience is One Vs All,
which is a transformation approach; it creates a model for each label through a
kernel classifier. We have chosen to use three different classifiers: logistic
regression, SVM and ANN. The results of our experimental study show that the
best accuracies are obtained using ANN model.

Original language	English
Title of host publication	Intelligent Systems Design and Applications (ISDA 2020)
Pages	478-487
Number of pages	10
DOIs	https://doi.org/10.1007/978-3-030-71187-0_44
Publication status	Published - 2020

Access to Document

10.1007/978-3-030-71187-0_44

Cite this

@inproceedings{fd269e37c60f4626bf765b65d858c43f,

title = "{"}One vs All{"} Classifier Analysis for Multi-label Movie Genre Classification Using Document Embedding.",

abstract = "The classification of movie genres from their synopses has attracted the attention of many researchers. Indeed, synopses are a source of relevant information that contributes to determinate movie genre. The automation of this classification process is very useful in several applications, such recommenda- tion systems. Moreover, movies can belong simultaneously to several genres (drama, action, comedy, horror), which reflects a typical problem of multi-label classification (MLC). In this article, we use a powerful representation of film synthesis via a document integration technique Doc2vec in the multi-label classification context. The technique used in our experience is One Vs All, which is a transformation approach; it creates a model for each label through a kernel classifier. We have chosen to use three different classifiers: logistic regression, SVM and ANN. The results of our experimental study show that the best accuracies are obtained using ANN model.",

author = "Sonia Guehria and Habiba Belleili and Nabiha Azizi and Belhaouari, \{Samir Brahim\}",

note = "DBLP License: DBLP's bibliographic metadata records provided through http://dblp.org/ are distributed under a Creative Commons CC0 1.0 Universal Public Domain Dedication. Although the bibliographic metadata records are provided consistent with CC0 1.0 Dedication, the content described by the metadata records is not. Content may be subject to copyright, rights of privacy, rights of publicity and other restrictions.",

year = "2020",

doi = "10.1007/978-3-030-71187-0\_44",

language = "English",

pages = "478--487",

booktitle = "Intelligent Systems Design and Applications (ISDA 2020)",

}

TY - GEN

T1 - "One vs All" Classifier Analysis for Multi-label Movie Genre Classification Using Document Embedding.

AU - Guehria, Sonia

AU - Belleili, Habiba

AU - Azizi, Nabiha

AU - Belhaouari, Samir Brahim

N1 - DBLP License: DBLP's bibliographic metadata records provided through http://dblp.org/ are distributed under a Creative Commons CC0 1.0 Universal Public Domain Dedication. Although the bibliographic metadata records are provided consistent with CC0 1.0 Dedication, the content described by the metadata records is not. Content may be subject to copyright, rights of privacy, rights of publicity and other restrictions.

PY - 2020

Y1 - 2020

N2 - The classification of movie genres from their synopses has attracted the attention of many researchers. Indeed, synopses are a source of relevant information that contributes to determinate movie genre. The automation of this classification process is very useful in several applications, such recommenda- tion systems. Moreover, movies can belong simultaneously to several genres (drama, action, comedy, horror), which reflects a typical problem of multi-label classification (MLC). In this article, we use a powerful representation of film synthesis via a document integration technique Doc2vec in the multi-label classification context. The technique used in our experience is One Vs All, which is a transformation approach; it creates a model for each label through a kernel classifier. We have chosen to use three different classifiers: logistic regression, SVM and ANN. The results of our experimental study show that the best accuracies are obtained using ANN model.

AB - The classification of movie genres from their synopses has attracted the attention of many researchers. Indeed, synopses are a source of relevant information that contributes to determinate movie genre. The automation of this classification process is very useful in several applications, such recommenda- tion systems. Moreover, movies can belong simultaneously to several genres (drama, action, comedy, horror), which reflects a typical problem of multi-label classification (MLC). In this article, we use a powerful representation of film synthesis via a document integration technique Doc2vec in the multi-label classification context. The technique used in our experience is One Vs All, which is a transformation approach; it creates a model for each label through a kernel classifier. We have chosen to use three different classifiers: logistic regression, SVM and ANN. The results of our experimental study show that the best accuracies are obtained using ANN model.

U2 - 10.1007/978-3-030-71187-0_44

DO - 10.1007/978-3-030-71187-0_44

M3 - Conference contribution

SP - 478

EP - 487

BT - Intelligent Systems Design and Applications (ISDA 2020)

ER -

"One vs All" Classifier Analysis for Multi-label Movie Genre Classification Using Document Embedding.

Abstract

Access to Document

Fingerprint

Cite this