Facebook activity event extraction system

Yuan Hao Lin; Chia Hui Chang

Facebook activity event extraction system

Yuan Hao Lin, Chia Hui Chang

Department of Computer Science and Information Engineering

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

2 Scopus citations

Abstract

The popularity of social networks has made them a perfect medium for activity or advertising campaign promotion. Therefore, many people use Facebook pages to announce their advertising campaign. The purpose of this study is to extract activity events by constructing two named entity recognition models, namely activity name and location, via a Web NER model generation tool [1]. We enhance the tool by improving the tokenizer and alignment technique. In addition, we also use a large database of FB checkin places for location name recognition improvement. For entity relation extraction, we apply sequential pattern mining to find rules for start date, end date, and location coupling. We use 1,300 posts from Facebook to test the activity event extraction performance. The experimental results show 0.727, 0.694 F₁score for activity name and location recognition; and 0.865, 0.72 F₁-score for start and end date extraction. Overall, the extraction performance for activity event extraction is 0.708.

Original language	English
Title of host publication	Proceedings of the 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016
Editors	Chung-Hsien Wu, Yuen-Hsien Tseng, Hung-Yu Kao, Lun-Wei Ku, Yu Tsao, Shih-Hung Wu
Publisher	The Association for Computational Linguistics and Chinese Language Processing (ACLCLP)
Pages	229-243
Number of pages	15
ISBN (Electronic)	9789573079293
State	Published - 1 Oct 2016
Event	28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016 - Tainan, Taiwan Duration: 6 Oct 2016 → 7 Oct 2016

Publication series

Name	Proceedings of the 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016

Conference

Conference	28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016
Country/Territory	Taiwan
City	Tainan
Period	6/10/16 → 7/10/16

Keywords

Activity Event Extraction
Named Entity Recognition
Social Media Event

Cite this

Lin, Y. H., & Chang, C. H. (2016). Facebook activity event extraction system. In C.-H. Wu, Y.-H. Tseng, H.-Y. Kao, L.-W. Ku, Y. Tsao, & S.-H. Wu (Eds.), Proceedings of the 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016 (pp. 229-243). (Proceedings of the 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016). The Association for Computational Linguistics and Chinese Language Processing (ACLCLP).

Lin, Yuan Hao ; Chang, Chia Hui. / Facebook activity event extraction system. Proceedings of the 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016. editor / Chung-Hsien Wu ; Yuen-Hsien Tseng ; Hung-Yu Kao ; Lun-Wei Ku ; Yu Tsao ; Shih-Hung Wu. The Association for Computational Linguistics and Chinese Language Processing (ACLCLP), 2016. pp. 229-243 (Proceedings of the 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016).

@inproceedings{74269eeb4de448fe9dc9e78a1a393d28,

title = "Facebook activity event extraction system",

abstract = "The popularity of social networks has made them a perfect medium for activity or advertising campaign promotion. Therefore, many people use Facebook pages to announce their advertising campaign. The purpose of this study is to extract activity events by constructing two named entity recognition models, namely activity name and location, via a Web NER model generation tool [1]. We enhance the tool by improving the tokenizer and alignment technique. In addition, we also use a large database of FB checkin places for location name recognition improvement. For entity relation extraction, we apply sequential pattern mining to find rules for start date, end date, and location coupling. We use 1,300 posts from Facebook to test the activity event extraction performance. The experimental results show 0.727, 0.694 F1score for activity name and location recognition; and 0.865, 0.72 F1-score for start and end date extraction. Overall, the extraction performance for activity event extraction is 0.708.",

keywords = "Activity Event Extraction, Named Entity Recognition, Social Media Event",

author = "Lin, {Yuan Hao} and Chang, {Chia Hui}",

note = "Publisher Copyright: {\textcopyright} Proceedings of the 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016.; 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016 ; Conference date: 06-10-2016 Through 07-10-2016",

year = "2016",

month = oct,

day = "1",

language = "???core.languages.en_GB???",

series = "Proceedings of the 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016",

publisher = "The Association for Computational Linguistics and Chinese Language Processing (ACLCLP)",

pages = "229--243",

editor = "Chung-Hsien Wu and Yuen-Hsien Tseng and Hung-Yu Kao and Lun-Wei Ku and Yu Tsao and Shih-Hung Wu",

booktitle = "Proceedings of the 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016",

}

Lin, YH & Chang, CH 2016, Facebook activity event extraction system. in C-H Wu, Y-H Tseng, H-Y Kao, L-W Ku, Y Tsao & S-H Wu (eds), Proceedings of the 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016. Proceedings of the 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016, The Association for Computational Linguistics and Chinese Language Processing (ACLCLP), pp. 229-243, 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016, Tainan, Taiwan, 6/10/16.

Facebook activity event extraction system. / Lin, Yuan Hao; Chang, Chia Hui.
Proceedings of the 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016. ed. / Chung-Hsien Wu; Yuen-Hsien Tseng; Hung-Yu Kao; Lun-Wei Ku; Yu Tsao; Shih-Hung Wu. The Association for Computational Linguistics and Chinese Language Processing (ACLCLP), 2016. p. 229-243 (Proceedings of the 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016).

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

TY - GEN

T1 - Facebook activity event extraction system

AU - Lin, Yuan Hao

AU - Chang, Chia Hui

N1 - Publisher Copyright: © Proceedings of the 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016.

PY - 2016/10/1

Y1 - 2016/10/1

N2 - The popularity of social networks has made them a perfect medium for activity or advertising campaign promotion. Therefore, many people use Facebook pages to announce their advertising campaign. The purpose of this study is to extract activity events by constructing two named entity recognition models, namely activity name and location, via a Web NER model generation tool [1]. We enhance the tool by improving the tokenizer and alignment technique. In addition, we also use a large database of FB checkin places for location name recognition improvement. For entity relation extraction, we apply sequential pattern mining to find rules for start date, end date, and location coupling. We use 1,300 posts from Facebook to test the activity event extraction performance. The experimental results show 0.727, 0.694 F1score for activity name and location recognition; and 0.865, 0.72 F1-score for start and end date extraction. Overall, the extraction performance for activity event extraction is 0.708.

AB - The popularity of social networks has made them a perfect medium for activity or advertising campaign promotion. Therefore, many people use Facebook pages to announce their advertising campaign. The purpose of this study is to extract activity events by constructing two named entity recognition models, namely activity name and location, via a Web NER model generation tool [1]. We enhance the tool by improving the tokenizer and alignment technique. In addition, we also use a large database of FB checkin places for location name recognition improvement. For entity relation extraction, we apply sequential pattern mining to find rules for start date, end date, and location coupling. We use 1,300 posts from Facebook to test the activity event extraction performance. The experimental results show 0.727, 0.694 F1score for activity name and location recognition; and 0.865, 0.72 F1-score for start and end date extraction. Overall, the extraction performance for activity event extraction is 0.708.

KW - Activity Event Extraction

KW - Named Entity Recognition

KW - Social Media Event

UR - http://www.scopus.com/inward/record.url?scp=85085939257&partnerID=8YFLogxK

M3 - 會議論文篇章

AN - SCOPUS:85085939257

T3 - Proceedings of the 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016

SP - 229

EP - 243

BT - Proceedings of the 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016

A2 - Wu, Chung-Hsien

A2 - Tseng, Yuen-Hsien

A2 - Kao, Hung-Yu

A2 - Ku, Lun-Wei

A2 - Tsao, Yu

A2 - Wu, Shih-Hung

PB - The Association for Computational Linguistics and Chinese Language Processing (ACLCLP)

T2 - 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016

Y2 - 6 October 2016 through 7 October 2016

ER -

Lin YH, Chang CH. Facebook activity event extraction system. In Wu CH, Tseng YH, Kao HY, Ku LW, Tsao Y, Wu SH, editors, Proceedings of the 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016. The Association for Computational Linguistics and Chinese Language Processing (ACLCLP). 2016. p. 229-243. (Proceedings of the 28th Conference on Computational Linguistics and Speech Processing, ROCLING 2016).

Facebook activity event extraction system

Abstract

Publication series

Conference

Keywords

Fingerprint

Cite this