Predictive team formation analysis via feature representation learning on social networks

Lo Pang Yun Ting, Cheng-Te Li, Kun-Ta Chuang

研究成果: Conference contribution

摘要

Team formation is to find a group of experts covering required skills and well collaborating together. Existing studies suffer from two defects: cannot afford flexible designation of team members and do not consider whether the formed team is truly adopted in practice. In this paper, we propose the Predictive Team Formation (PTF) problem. PTF provides the flexibility of designated members and delivers the prediction-based formulation to compose the team. We propose two methods by learning the feature representations of experts based on node2vec [4]. One is Biased-n2v that models the topic bias of each expert in the social network. The other is Guided-n2v that refines the transition probabilities between skills and experts to guide the random walk in a heterogeneous graph of expert-expert, expert-skill, and skill-skill. Experiments conducted on DBLP and IMDb datasets exhibit that our methods can significantly outperform the state-of-the-art optimization-based approaches in terms of prediction recall. We also reveal that the designated members with tight social connections can lead to better performance.

原文English
主出版物標題Advances in Knowledge Discovery and Data Mining - 22nd Pacific-Asia Conference, PAKDD 2018, Proceedings
編輯Geoffrey I. Webb, Dinh Phung, Mohadeseh Ganji, Lida Rashidi, Vincent S. Tseng, Bao Ho
發行者Springer Verlag
頁面790-802
頁數13
ISBN(列印)9783319930398
DOIs
出版狀態Published - 2018 一月 1
事件22nd Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining, PAKDD 2018 - Melbourne, Australia
持續時間: 2018 六月 32018 六月 6

出版系列

名字Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
10939 LNAI
ISSN(列印)0302-9743
ISSN(電子)1611-3349

Other

Other22nd Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining, PAKDD 2018
國家Australia
城市Melbourne
期間18-06-0318-06-06

指紋

Social Networks
Defects
Experiments
Prediction
Learning
Transition Probability
Biased
Random walk
Covering
Flexibility
Skills
Optimization
Formulation
Graph in graph theory
Experiment

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

引用此文

Ting, L. P. Y., Li, C-T., & Chuang, K-T. (2018). Predictive team formation analysis via feature representation learning on social networks. 於 G. I. Webb, D. Phung, M. Ganji, L. Rashidi, V. S. Tseng, & B. Ho (編輯), Advances in Knowledge Discovery and Data Mining - 22nd Pacific-Asia Conference, PAKDD 2018, Proceedings (頁 790-802). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); 卷 10939 LNAI). Springer Verlag. https://doi.org/10.1007/978-3-319-93040-4_62
Ting, Lo Pang Yun ; Li, Cheng-Te ; Chuang, Kun-Ta. / Predictive team formation analysis via feature representation learning on social networks. Advances in Knowledge Discovery and Data Mining - 22nd Pacific-Asia Conference, PAKDD 2018, Proceedings. 編輯 / Geoffrey I. Webb ; Dinh Phung ; Mohadeseh Ganji ; Lida Rashidi ; Vincent S. Tseng ; Bao Ho. Springer Verlag, 2018. 頁 790-802 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{9b3ae1e754e84f3b924d8a2b0a22fb17,
title = "Predictive team formation analysis via feature representation learning on social networks",
abstract = "Team formation is to find a group of experts covering required skills and well collaborating together. Existing studies suffer from two defects: cannot afford flexible designation of team members and do not consider whether the formed team is truly adopted in practice. In this paper, we propose the Predictive Team Formation (PTF) problem. PTF provides the flexibility of designated members and delivers the prediction-based formulation to compose the team. We propose two methods by learning the feature representations of experts based on node2vec [4]. One is Biased-n2v that models the topic bias of each expert in the social network. The other is Guided-n2v that refines the transition probabilities between skills and experts to guide the random walk in a heterogeneous graph of expert-expert, expert-skill, and skill-skill. Experiments conducted on DBLP and IMDb datasets exhibit that our methods can significantly outperform the state-of-the-art optimization-based approaches in terms of prediction recall. We also reveal that the designated members with tight social connections can lead to better performance.",
author = "Ting, {Lo Pang Yun} and Cheng-Te Li and Kun-Ta Chuang",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/978-3-319-93040-4_62",
language = "English",
isbn = "9783319930398",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "790--802",
editor = "Webb, {Geoffrey I.} and Dinh Phung and Mohadeseh Ganji and Lida Rashidi and Tseng, {Vincent S.} and Bao Ho",
booktitle = "Advances in Knowledge Discovery and Data Mining - 22nd Pacific-Asia Conference, PAKDD 2018, Proceedings",
address = "Germany",

}

Ting, LPY, Li, C-T & Chuang, K-T 2018, Predictive team formation analysis via feature representation learning on social networks. 於 GI Webb, D Phung, M Ganji, L Rashidi, VS Tseng & B Ho (編輯), Advances in Knowledge Discovery and Data Mining - 22nd Pacific-Asia Conference, PAKDD 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 卷 10939 LNAI, Springer Verlag, 頁 790-802, 22nd Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining, PAKDD 2018, Melbourne, Australia, 18-06-03. https://doi.org/10.1007/978-3-319-93040-4_62

Predictive team formation analysis via feature representation learning on social networks. / Ting, Lo Pang Yun; Li, Cheng-Te; Chuang, Kun-Ta.

Advances in Knowledge Discovery and Data Mining - 22nd Pacific-Asia Conference, PAKDD 2018, Proceedings. 編輯 / Geoffrey I. Webb; Dinh Phung; Mohadeseh Ganji; Lida Rashidi; Vincent S. Tseng; Bao Ho. Springer Verlag, 2018. p. 790-802 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); 卷 10939 LNAI).

研究成果: Conference contribution

TY - GEN

T1 - Predictive team formation analysis via feature representation learning on social networks

AU - Ting, Lo Pang Yun

AU - Li, Cheng-Te

AU - Chuang, Kun-Ta

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Team formation is to find a group of experts covering required skills and well collaborating together. Existing studies suffer from two defects: cannot afford flexible designation of team members and do not consider whether the formed team is truly adopted in practice. In this paper, we propose the Predictive Team Formation (PTF) problem. PTF provides the flexibility of designated members and delivers the prediction-based formulation to compose the team. We propose two methods by learning the feature representations of experts based on node2vec [4]. One is Biased-n2v that models the topic bias of each expert in the social network. The other is Guided-n2v that refines the transition probabilities between skills and experts to guide the random walk in a heterogeneous graph of expert-expert, expert-skill, and skill-skill. Experiments conducted on DBLP and IMDb datasets exhibit that our methods can significantly outperform the state-of-the-art optimization-based approaches in terms of prediction recall. We also reveal that the designated members with tight social connections can lead to better performance.

AB - Team formation is to find a group of experts covering required skills and well collaborating together. Existing studies suffer from two defects: cannot afford flexible designation of team members and do not consider whether the formed team is truly adopted in practice. In this paper, we propose the Predictive Team Formation (PTF) problem. PTF provides the flexibility of designated members and delivers the prediction-based formulation to compose the team. We propose two methods by learning the feature representations of experts based on node2vec [4]. One is Biased-n2v that models the topic bias of each expert in the social network. The other is Guided-n2v that refines the transition probabilities between skills and experts to guide the random walk in a heterogeneous graph of expert-expert, expert-skill, and skill-skill. Experiments conducted on DBLP and IMDb datasets exhibit that our methods can significantly outperform the state-of-the-art optimization-based approaches in terms of prediction recall. We also reveal that the designated members with tight social connections can lead to better performance.

UR - http://www.scopus.com/inward/record.url?scp=85049363230&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85049363230&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-93040-4_62

DO - 10.1007/978-3-319-93040-4_62

M3 - Conference contribution

AN - SCOPUS:85049363230

SN - 9783319930398

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 790

EP - 802

BT - Advances in Knowledge Discovery and Data Mining - 22nd Pacific-Asia Conference, PAKDD 2018, Proceedings

A2 - Webb, Geoffrey I.

A2 - Phung, Dinh

A2 - Ganji, Mohadeseh

A2 - Rashidi, Lida

A2 - Tseng, Vincent S.

A2 - Ho, Bao

PB - Springer Verlag

ER -

Ting LPY, Li C-T, Chuang K-T. Predictive team formation analysis via feature representation learning on social networks. 於 Webb GI, Phung D, Ganji M, Rashidi L, Tseng VS, Ho B, 編輯, Advances in Knowledge Discovery and Data Mining - 22nd Pacific-Asia Conference, PAKDD 2018, Proceedings. Springer Verlag. 2018. p. 790-802. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-93040-4_62