News recommendations using CF-IDF

A.C. Hogenboom; F. Frasincar; U. Kaymak; F.M.G. Jong, de

News recommendations using CF-IDF

A.C. Hogenboom, F. Frasincar, U. Kaymak, F.M.G. Jong, de

Information Systems IE&IS

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic

170 Downloads (Pure)

Abstract

Most of the traditional recommendation algorithms are based on TF-IDF, a term-based weighting method. This paper proposes a new method for recommending news items based on the weighting of the occurrences of references to concepts, which we call Concept Frequency-Inverse Document Frequency (CFIDF). In an experimental setup we apply CF-IDF to a set of newswires in which we detect 1; 167 instances of a set of 65 concepts from a domain ontology. The proposed method yields significantly better results with respect to accuracy, recall, and F1 than the TF-IDF method we use as a basis for comparison.

Original language	English
Title of host publication	Proceedings of the 23rd Benelux Conference on Artificial Intelligence (BNAIC 2011), November 3-4, 2011, Gent, Belgium
Editors	P. De Causmaecker, J. Maervoet, T. Messelis, K. Verbeeck, T. Vermeulen
Place of Publication	Gent
Publisher	BNAIC
Pages	397-398
Publication status	Published - 2011
Event	23rd Benelux Conference on Artificial Intelligence (BNAIC 2011) - Gent, Belgium Duration: 3 Nov 2011 → 4 Nov 2011 Conference number: 23

Conference

Conference	23rd Benelux Conference on Artificial Intelligence (BNAIC 2011)
Abbreviated title	BNAIC 2011
Country/Territory	Belgium
City	Gent
Period	3/11/11 → 4/11/11

Access to Document

publishers versionFinal published version, 111 KB

Cite this

@inproceedings{161f239dae144e068d9f4e47a9937f79,

title = "News recommendations using CF-IDF",

abstract = "Most of the traditional recommendation algorithms are based on TF-IDF, a term-based weighting method. This paper proposes a new method for recommending news items based on the weighting of the occurrences of references to concepts, which we call Concept Frequency-Inverse Document Frequency (CFIDF). In an experimental setup we apply CF-IDF to a set of newswires in which we detect 1; 167 instances of a set of 65 concepts from a domain ontology. The proposed method yields significantly better results with respect to accuracy, recall, and F1 than the TF-IDF method we use as a basis for comparison.",

author = "A.C. Hogenboom and F. Frasincar and U. Kaymak and {Jong, de}, F.M.G.",

year = "2011",

language = "English",

pages = "397--398",

editor = "{De Causmaecker}, P. and J. Maervoet and T. Messelis and K. Verbeeck and T. Vermeulen",

booktitle = "Proceedings of the 23rd Benelux Conference on Artificial Intelligence (BNAIC 2011), November 3-4, 2011, Gent, Belgium",

publisher = "BNAIC",

note = "23rd Benelux Conference on Artificial Intelligence (BNAIC 2011), BNAIC 2011 ; Conference date: 03-11-2011 Through 04-11-2011",

}

Hogenboom, AC, Frasincar, F, Kaymak, U & Jong, de, FMG 2011, News recommendations using CF-IDF. in P De Causmaecker, J Maervoet, T Messelis, K Verbeeck & T Vermeulen (eds), Proceedings of the 23rd Benelux Conference on Artificial Intelligence (BNAIC 2011), November 3-4, 2011, Gent, Belgium. BNAIC, Gent, pp. 397-398, 23rd Benelux Conference on Artificial Intelligence (BNAIC 2011), Gent, Belgium, 3/11/11.

News recommendations using CF-IDF. / Hogenboom, A.C.; Frasincar, F.; Kaymak, U. et al.
Proceedings of the 23rd Benelux Conference on Artificial Intelligence (BNAIC 2011), November 3-4, 2011, Gent, Belgium. ed. / P. De Causmaecker; J. Maervoet; T. Messelis; K. Verbeeck; T. Vermeulen. Gent: BNAIC, 2011. p. 397-398.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic

TY - GEN

T1 - News recommendations using CF-IDF

AU - Hogenboom, A.C.

AU - Frasincar, F.

AU - Kaymak, U.

AU - Jong, de, F.M.G.

N1 - Conference code: 23

PY - 2011

Y1 - 2011

N2 - Most of the traditional recommendation algorithms are based on TF-IDF, a term-based weighting method. This paper proposes a new method for recommending news items based on the weighting of the occurrences of references to concepts, which we call Concept Frequency-Inverse Document Frequency (CFIDF). In an experimental setup we apply CF-IDF to a set of newswires in which we detect 1; 167 instances of a set of 65 concepts from a domain ontology. The proposed method yields significantly better results with respect to accuracy, recall, and F1 than the TF-IDF method we use as a basis for comparison.

AB - Most of the traditional recommendation algorithms are based on TF-IDF, a term-based weighting method. This paper proposes a new method for recommending news items based on the weighting of the occurrences of references to concepts, which we call Concept Frequency-Inverse Document Frequency (CFIDF). In an experimental setup we apply CF-IDF to a set of newswires in which we detect 1; 167 instances of a set of 65 concepts from a domain ontology. The proposed method yields significantly better results with respect to accuracy, recall, and F1 than the TF-IDF method we use as a basis for comparison.

M3 - Conference contribution

SP - 397

EP - 398

BT - Proceedings of the 23rd Benelux Conference on Artificial Intelligence (BNAIC 2011), November 3-4, 2011, Gent, Belgium

A2 - De Causmaecker, P.

A2 - Maervoet, J.

A2 - Messelis, T.

A2 - Verbeeck, K.

A2 - Vermeulen, T.

PB - BNAIC

CY - Gent

T2 - 23rd Benelux Conference on Artificial Intelligence (BNAIC 2011)

Y2 - 3 November 2011 through 4 November 2011

ER -

News recommendations using CF-IDF

Abstract

Conference

Access to Document

Fingerprint

Cite this